2 /* $NetBSD: msdosfs_conv.c,v 1.25 1997/11/17 15:36:40 ws Exp $ */
5 * Copyright (C) 1995, 1997 Wolfgang Solfrank.
6 * Copyright (C) 1995, 1997 TooLs GmbH.
8 * Original code by Paul Popelka (paulp@uts.amdahl.com) (see below).
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 3. All advertising materials mentioning features or use of this software
19 * must display the following acknowledgement:
20 * This product includes software developed by TooLs GmbH.
21 * 4. The name of TooLs GmbH may not be used to endorse or promote products
22 * derived from this software without specific prior written permission.
24 * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR
25 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
26 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
27 * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
28 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
29 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
30 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
31 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
32 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
33 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
36 * Written by Paul Popelka (paulp@uts.amdahl.com)
38 * You can do anything you want with this software, just don't say you wrote
39 * it, and don't remove this notice.
41 * This software is provided "as is".
43 * The author supplies this software to be publicly redistributed on the
44 * understanding that the author is not responsible for the correct
45 * functioning of this software in any circumstances and is not liable for
46 * any damages caused by this software.
51 #include <sys/param.h>
52 #include <sys/systm.h>
53 #include <sys/dirent.h>
54 #include <sys/iconv.h>
55 #include <sys/malloc.h>
56 #include <sys/mount.h>
58 #include <fs/msdosfs/bpb.h>
59 #include <fs/msdosfs/direntry.h>
60 #include <fs/msdosfs/msdosfsmount.h>
62 extern struct iconv_functions *msdosfs_iconv;
64 static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle);
65 static u_int16_t dos2unixchr(const u_char **, size_t *, int, struct msdosfsmount *);
66 static u_int16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *);
67 static u_int16_t win2unixchr(u_int16_t, struct msdosfsmount *);
68 static u_int16_t unix2winchr(const u_char **, size_t *, int, struct msdosfsmount *);
70 static char *nambuf_ptr;
71 static size_t nambuf_len;
72 static int nambuf_last_id;
75 * 0 - character disallowed in long file name.
76 * 1 - character should be replaced by '_' in DOS file name,
77 * and generation number inserted.
78 * 2 - character ('.' and ' ') should be skipped in DOS file name,
79 * and generation number inserted.
83 /* iso8859-1 -> cp850 */
84 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */
85 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */
86 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */
87 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */
88 2, 0x21, 0, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
89 0x28, 0x29, 0, 1, 1, 0x2d, 2, 0, /* 28-2f */
90 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
91 0x38, 0x39, 0, 1, 0, 1, 0, 0, /* 38-3f */
92 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */
93 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */
94 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */
95 0x58, 0x59, 0x5a, 1, 0, 1, 0x5e, 0x5f, /* 58-5f */
96 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 60-67 */
97 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 68-6f */
98 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 70-77 */
99 0x58, 0x59, 0x5a, 0x7b, 0, 0x7d, 0x7e, 0, /* 78-7f */
100 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */
101 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */
102 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */
103 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */
104 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */
105 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */
106 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */
107 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */
108 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */
109 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */
110 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */
111 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */
112 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */
113 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */
114 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */
115 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */
120 /* cp850 -> iso8859-1 */
121 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 00-07 */
122 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 08-0f */
123 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 10-17 */
124 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 18-1f */
125 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
126 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
127 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
128 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
129 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */
130 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */
131 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */
132 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
133 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
134 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
135 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
136 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
137 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */
138 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */
139 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */
140 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, 0x3f, /* 98-9f */
141 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */
142 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */
143 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xc1, 0xc2, 0xc0, /* b0-b7 */
144 0xa9, 0x3f, 0x3f, 0x3f, 0x3f, 0xa2, 0xa5, 0x3f, /* b8-bf */
145 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xe3, 0xc3, /* c0-c7 */
146 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xa4, /* c8-cf */
147 0xf0, 0xd0, 0xca, 0xcb, 0xc8, 0x3f, 0xcd, 0xce, /* d0-d7 */
148 0xcf, 0x3f, 0x3f, 0x3f, 0x3f, 0xa6, 0xcc, 0x3f, /* d8-df */
149 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */
150 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */
151 0xad, 0xb1, 0x3f, 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */
152 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, 0x3f, 0x3f, /* f8-ff */
158 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */
159 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */
160 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */
161 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */
162 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
163 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
164 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
165 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
166 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */
167 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */
168 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */
169 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
170 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
171 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
172 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
173 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
174 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */
175 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */
176 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */
177 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */
178 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */
179 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */
180 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */
181 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */
182 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */
183 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */
184 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */
185 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */
186 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */
187 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */
188 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */
189 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */
195 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */
196 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */
197 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */
198 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */
199 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
200 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
201 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
202 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
203 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */
204 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */
205 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */
206 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
207 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
208 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
209 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
210 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
211 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */
212 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */
213 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */
214 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */
215 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */
216 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */
217 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */
218 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */
219 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */
220 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */
221 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */
222 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */
223 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */
224 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */
225 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */
226 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */
230 * DOS filenames are made of 2 parts, the name part and the extension part.
231 * The name part is 8 characters long and the extension part is 3
232 * characters long. They may contain trailing blanks if the name or
233 * extension are not long enough to fill their respective fields.
237 * Convert a DOS filename to a unix filename. And, return the number of
238 * characters in the resulting unix filename excluding the terminating
242 dos2unixfn(dn, un, lower, pmp)
246 struct msdosfsmount *pmp;
253 * If first char of the filename is SLOT_E5 (0x05), then the real
254 * first char of the filename should be 0xe5. But, they couldn't
255 * just have a 0xe5 mean 0xe5 because that is used to mean a freed
256 * directory slot. Another dos quirk.
262 * Copy the name portion into the unix filename string.
264 for (i = 8; i > 0 && *dn != ' ';) {
265 c = dos2unixchr((const u_char **)&dn, &i, lower & LCASE_BASE,
277 * Now, if there is an extension then put in a period and copy in
283 for (i = 3; i > 0 && *dn != ' ';) {
284 c = dos2unixchr((const u_char **)&dn, &i,
285 lower & LCASE_EXT, pmp);
300 * Convert a unix filename to a DOS filename according to Win95 rules.
301 * If applicable and gen is not 0, it is inserted into the converted
302 * filename as a generation number.
304 * 0 if name couldn't be converted
305 * 1 if the converted name is the same as the original
306 * (no long filename entry necessary for Win95)
307 * 2 if conversion was successful
308 * 3 if conversion was successful and generation number was inserted
311 unix2dosfn(un, dn, unlen, gen, pmp)
316 struct msdosfsmount *pmp;
321 const u_char *cp, *dp, *dp1;
322 u_char gentext[6], *wcp;
326 * Fill the dos filename string with blanks. These are DOS's pad
329 for (i = 0; i < 11; i++)
334 * The filenames "." and ".." are handled specially, since they
335 * don't follow dos filename rules.
337 if (un[0] == '.' && unlen == 1) {
341 if (un[0] == '.' && un[1] == '.' && unlen == 2) {
348 * Filenames with only blanks and dots are not allowed!
350 for (cp = un, i = unlen; --i >= 0; cp++)
351 if (*cp != ' ' && *cp != '.')
358 * Filenames with some characters are not allowed!
360 for (cp = un, i = unlen; i > 0;)
361 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0)
365 * Now find the extension
366 * Note: dot as first char doesn't start extension
367 * and trailing dots and blanks are ignored
368 * Note(2003/7): It seems recent Windows has
369 * defferent rule than this code, that Windows
370 * ignores all dots before extension, and use all
371 * chars as filename except for dots.
374 for (cp = un + 1, i = unlen - 1; --i >= 0;) {
391 * Now convert it (this part is for extension).
392 * As Windows XP do, if it's not ascii char,
393 * this function should return 2 or 3, so that checkng out Unicode name.
399 l = unlen - (dp - un);
400 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) {
401 c = unix2doschr(&cp, (size_t *)&i, pmp);
417 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3)
432 for (dp = cp; *--dp == ' ' || *dp == '.';);
437 * Now convert the rest of the name
439 for (i = dp - un, j = 0; un < dp && j < 8; j++) {
440 c = unix2doschr(&un, &i, pmp);
456 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3)
470 * If we didn't have any chars in filename,
477 * If there wasn't any char dropped,
478 * there is no place for generation numbers
487 * Now insert the generation number into the filename part
491 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10)
492 *--wcp = gen % 10 + '0';
497 for (i = 8; dn[--i] == ' ';);
499 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i)
500 i = 8 - (gentext + sizeof(gentext) - wcp + 1);
502 * Correct posision to where insert the generation number
505 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u);
508 while (wcp < gentext + sizeof(gentext))
512 * Tail of the filename should be space
520 * The first character cannot be E5,
521 * because that means a deleted entry
530 * Create a Win95 long name directory entry
531 * Note: assumes that the filename is valid,
532 * i.e. doesn't consist solely of blanks and dots
535 unix2winfn(un, unlen, wep, cnt, chksum, pmp)
538 struct winentry *wep;
541 struct msdosfsmount *pmp;
548 * Drop trailing blanks and dots
550 unlen = winLenFixup(un, unlen);
553 * Cut *un for this slot
555 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2,
556 pmp->pm_flags, pmp->pm_u2w);
559 * Initialize winentry to some useful default
561 for (wcp = (u_int8_t *)wep, i = sizeof(*wep); --i >= 0; *wcp++ = 0xff);
563 wep->weAttributes = ATTR_WIN95;
564 wep->weReserved1 = 0;
565 wep->weChksum = chksum;
566 wep->weReserved2 = 0;
569 * Now convert the filename parts
572 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) {
573 code = unix2winchr(&un, &unlen, 0, pmp);
579 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) {
580 code = unix2winchr(&un, &unlen, 0, pmp);
586 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) {
587 code = unix2winchr(&un, &unlen, 0, pmp);
600 * Compare our filename to the one in the Win95 entry
601 * Returns the checksum or -1 if no match
604 winChkName(un, unlen, chksum, pmp)
608 struct msdosfsmount *pmp;
613 struct dirent dirbuf;
616 * We alread have winentry in mbnambuf
618 if (!mbnambuf_flush(&dirbuf) || !dirbuf.d_namlen)
622 printf("winChkName(): un=%s:%d,d_name=%s:%d\n", un, unlen,
628 * Compare the name parts
630 len = dirbuf.d_namlen;
634 for (np = dirbuf.d_name; unlen > 0 && len > 0;) {
636 * Comparison must be case insensitive, because FAT disallows
637 * to look up or create files in case sensitive even when
638 * it's a long file name.
640 c1 = unix2winchr((const u_char **)&np, &len, LCASE_BASE, pmp);
641 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp);
649 * Convert Win95 filename to dirbuf.
650 * Returns the checksum or -1 if impossible
653 win2unixfn(wep, chksum, pmp)
654 struct winentry *wep;
656 struct msdosfsmount *pmp;
659 u_int8_t *np, name[WIN_CHARS * 2 + 1];
663 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS)
664 || !(wep->weCnt&WIN_CNT))
668 * First compare checksums
670 if (wep->weCnt&WIN_LAST) {
671 chksum = wep->weChksum;
672 } else if (chksum != wep->weChksum)
678 * Convert the name parts
681 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) {
682 code = (cp[1] << 8) | cp[0];
686 mbnambuf_write(name, (wep->weCnt & WIN_CNT) - 1);
692 code = win2unixchr(code, pmp);
700 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) {
701 code = (cp[1] << 8) | cp[0];
705 mbnambuf_write(name, (wep->weCnt & WIN_CNT) - 1);
711 code = win2unixchr(code, pmp);
719 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) {
720 code = (cp[1] << 8) | cp[0];
724 mbnambuf_write(name, (wep->weCnt & WIN_CNT) - 1);
730 code = win2unixchr(code, pmp);
739 mbnambuf_write(name, (wep->weCnt & WIN_CNT) - 1);
744 * Compute the unrolled checksum of a DOS filename for Win95 LFN use.
747 winChksum(struct direntry *dep)
752 s = ((s << 7) | (s >> 1)) + dep->deName[1];
753 s = ((s << 7) | (s >> 1)) + dep->deName[2];
754 s = ((s << 7) | (s >> 1)) + dep->deName[3];
755 s = ((s << 7) | (s >> 1)) + dep->deName[4];
756 s = ((s << 7) | (s >> 1)) + dep->deName[5];
757 s = ((s << 7) | (s >> 1)) + dep->deName[6];
758 s = ((s << 7) | (s >> 1)) + dep->deName[7];
759 s = ((s << 7) | (s >> 1)) + dep->deExtension[0];
760 s = ((s << 7) | (s >> 1)) + dep->deExtension[1];
761 s = ((s << 7) | (s >> 1)) + dep->deExtension[2];
767 * Determine the number of slots necessary for Win95 names
770 winSlotCnt(un, unlen, pmp)
773 struct msdosfsmount *pmp;
776 char wn[WIN_MAXLEN * 2 + 1], *wnp;
778 unlen = winLenFixup(un, unlen);
780 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
781 wlen = WIN_MAXLEN * 2;
783 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen);
786 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS);
789 if (unlen > WIN_MAXLEN)
791 return howmany(unlen, WIN_CHARS);
795 * Determine the number of bytes neccesary for Win95 names
798 winLenFixup(un, unlen)
802 for (un += unlen; unlen > 0; unlen--)
803 if (*--un != ' ' && *un != '.')
809 * Store an area with multi byte string instr, and reterns left
810 * byte of instr and moves pointer forward. The area's size is
814 mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle)
816 char *outp, outstr[outlen * weight + 1];
818 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) {
821 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen);
825 (*instr) += min(inlen, outlen);
826 return (inlen - min(inlen, outlen));
830 * Convert DOS char to Local char
833 dos2unixchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp)
836 char *outp, outbuf[3];
840 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
844 if (lower & (LCASE_BASE | LCASE_EXT))
845 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr,
846 ilen, &outp, &olen, KICONV_LOWER);
848 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr,
853 * return '?' if failed to convert
863 wc |= (*(outp - len - 1) & 0xff) << (len << 3);
870 if (lower & (LCASE_BASE | LCASE_EXT))
872 return ((u_int16_t)c);
876 * Convert Local char to DOS char
879 unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp)
882 char *up, *outp, unicode[3], outbuf[3];
884 size_t len, ucslen, unixlen, olen;
886 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
888 * to hide an invisible character, using a unicode filter
893 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr,
895 unixlen = len - *ilen;
898 * cannot be converted
907 * return magic number for ascii char
919 * now convert using libiconv
926 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr,
927 ilen, &outp, &olen, KICONV_FROM_UPPER);
931 * cannot be converted, but has unicode char, should return magic number
941 wc |= (*(outp - len - 1) & 0xff) << (len << 3);
949 return ((u_int16_t)c);
953 * Convert Windows char to Local char
956 win2unixchr(u_int16_t wc, struct msdosfsmount *pmp)
958 u_char *inp, *outp, inbuf[3], outbuf[3];
959 size_t ilen, olen, len;
964 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
965 inbuf[0] = (u_char)(wc>>8);
966 inbuf[1] = (u_char)wc;
969 ilen = olen = len = 2;
972 msdosfs_iconv->convchr(pmp->pm_w2u, (const char **)&inp, &ilen,
973 (char **)&outp, &olen);
977 * return '?' if failed to convert
986 wc |= (*(outp - len - 1) & 0xff) << (len << 3);
997 * Convert Local char to Windows char
1000 unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp)
1002 u_char *outp, outbuf[3];
1009 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
1012 if (lower & (LCASE_BASE | LCASE_EXT))
1013 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr,
1014 ilen, (char **)&outp, &olen,
1017 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr,
1018 ilen, (char **)&outp, &olen);
1021 * return '0' if end of filename
1026 wc = (outbuf[0]<<8) | outbuf[1];
1033 if (lower & (LCASE_BASE | LCASE_EXT))
1040 * Initialize the temporary concatenation buffer (once) and mark it as
1041 * empty on subsequent calls.
1047 if (nambuf_ptr == NULL) {
1048 nambuf_ptr = malloc(MAXNAMLEN + 1, M_MSDOSFSMNT, M_WAITOK);
1049 nambuf_ptr[MAXNAMLEN] = '\0';
1052 nambuf_last_id = -1;
1056 * Fill out our concatenation buffer with the given substring, at the offset
1057 * specified by its id. Since this function must be called with ids in
1058 * descending order, we take advantage of the fact that ASCII substrings are
1059 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all
1060 * previous (i.e. higher id) substrings upwards to make room for this one.
1061 * This only penalizes portions of substrings that contain more than
1062 * WIN_CHARS bytes when they are first encountered.
1065 mbnambuf_write(char *name, int id)
1070 KASSERT(nambuf_len == 0 || id == nambuf_last_id - 1,
1071 ("non-decreasing id, id %d last id %d", id, nambuf_last_id));
1073 /* Store this substring in a WIN_CHAR-aligned slot. */
1074 slot = nambuf_ptr + (id * WIN_CHARS);
1075 count = strlen(name);
1076 if (nambuf_len + count > MAXNAMLEN) {
1077 printf("msdosfs: file name %zu too long\n", nambuf_len + count);
1081 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */
1082 if (count > WIN_CHARS && nambuf_len != 0)
1083 bcopy(slot + WIN_CHARS, slot + count, nambuf_len);
1085 /* Copy in the substring to its slot and update length so far. */
1086 bcopy(name, slot, count);
1087 nambuf_len += count;
1088 nambuf_last_id = id;
1092 * Take the completed string and use it to setup the struct dirent.
1093 * Be sure to always nul-terminate the d_name and then copy the string
1094 * from our buffer. Note that this function assumes the full string has
1095 * been reassembled in the buffer. If it's called before all substrings
1096 * have been written via mbnambuf_write(), the result will be incorrect.
1099 mbnambuf_flush(struct dirent *dp)
1102 if (nambuf_len > sizeof(dp->d_name) - 1) {
1106 bcopy(nambuf_ptr, dp->d_name, nambuf_len);
1107 dp->d_name[nambuf_len] = '\0';
1108 dp->d_namlen = nambuf_len;
1111 return (dp->d_name);