1 | /* $NetBSD: msdosfs_conv.c,v 1.17 2016/06/30 09:34:01 nonaka Exp $ */ |
2 | |
3 | /*- |
4 | * Copyright (C) 1995, 1997 Wolfgang Solfrank. |
5 | * Copyright (C) 1995, 1997 TooLs GmbH. |
6 | * All rights reserved. |
7 | * Original code by Paul Popelka (paulp@uts.amdahl.com) (see below). |
8 | * |
9 | * Redistribution and use in source and binary forms, with or without |
10 | * modification, are permitted provided that the following conditions |
11 | * are met: |
12 | * 1. Redistributions of source code must retain the above copyright |
13 | * notice, this list of conditions and the following disclaimer. |
14 | * 2. Redistributions in binary form must reproduce the above copyright |
15 | * notice, this list of conditions and the following disclaimer in the |
16 | * documentation and/or other materials provided with the distribution. |
17 | * 3. All advertising materials mentioning features or use of this software |
18 | * must display the following acknowledgement: |
19 | * This product includes software developed by TooLs GmbH. |
20 | * 4. The name of TooLs GmbH may not be used to endorse or promote products |
21 | * derived from this software without specific prior written permission. |
22 | * |
23 | * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR |
24 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES |
25 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. |
26 | * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
27 | * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
28 | * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; |
29 | * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, |
30 | * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR |
31 | * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF |
32 | * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
33 | */ |
34 | /* |
35 | * Written by Paul Popelka (paulp@uts.amdahl.com) |
36 | * |
37 | * You can do anything you want with this software, just don't say you wrote |
38 | * it, and don't remove this notice. |
39 | * |
40 | * This software is provided "as is". |
41 | * |
42 | * The author supplies this software to be publicly redistributed on the |
43 | * understanding that the author is not responsible for the correct |
44 | * functioning of this software in any circumstances and is not liable for |
45 | * any damages caused by this software. |
46 | * |
47 | * October 1992 |
48 | * |
49 | */ |
50 | |
51 | #if HAVE_NBTOOL_CONFIG_H |
52 | #include "nbtool_config.h" |
53 | #endif |
54 | |
55 | #ifndef _KERNEL |
56 | #include <assert.h> |
57 | #define KASSERT(x) assert(x) |
58 | #endif |
59 | |
60 | #include <sys/cdefs.h> |
61 | __KERNEL_RCSID(0, "$NetBSD: msdosfs_conv.c,v 1.17 2016/06/30 09:34:01 nonaka Exp $" ); |
62 | |
63 | /* |
64 | * System include files. |
65 | */ |
66 | #include <sys/param.h> |
67 | #include <sys/time.h> |
68 | #include <sys/endian.h> |
69 | #ifdef _KERNEL |
70 | #include <sys/dirent.h> |
71 | #include <sys/systm.h> |
72 | #include <sys/kernel.h> |
73 | #include <sys/vnode.h> |
74 | #else |
75 | #include <stdio.h> |
76 | #include <string.h> |
77 | #include <dirent.h> |
78 | #include <sys/queue.h> |
79 | #endif |
80 | #include <dev/clock_subr.h> |
81 | |
82 | /* |
83 | * MSDOSFS include files. |
84 | */ |
85 | #include <fs/msdosfs/direntry.h> |
86 | #include <fs/msdosfs/denode.h> |
87 | |
88 | static int invalidname(const u_int16_t *, int); |
89 | |
90 | static int ucs2utf8(const u_int16_t *, u_int8_t *, int); |
91 | static int utf8ucs2(const u_int8_t *, int, u_int16_t *); |
92 | |
93 | static int ucs2utf8str(const u_int16_t *, int, u_int8_t *, int); |
94 | static int utf8ucs2str(const u_int8_t *, int, u_int16_t *, int); |
95 | static int ucs2char8str(const u_int16_t *, int, u_int8_t *, int); |
96 | static int char8ucs2str(const u_int8_t *, int, u_int16_t *, int); |
97 | |
98 | static void ucs2pad(u_int16_t *, int, int); |
99 | |
100 | static u_int16_t ucs2fold(u_int16_t); |
101 | static int ucs2match(u_int16_t *, u_int16_t *, int n); |
102 | static int char8match(u_int16_t *, u_int16_t *, int n); |
103 | |
104 | /* |
105 | * The number of seconds between Jan 1, 1970 and Jan 1, 1980. In that |
106 | * interval there were 8 regular years and 2 leap years. |
107 | */ |
108 | #define DOSBIASYEAR 1980 |
109 | #define SECONDSTO1980 (((8 * 365) + (2 * 366)) * (24 * 60 * 60)) |
110 | /* |
111 | * msdos fs can not store dates beyound the year 2234 |
112 | */ |
113 | #define DOSMAXYEAR ((DD_YEAR_MASK >> DD_YEAR_SHIFT) + DOSBIASYEAR) |
114 | |
115 | /* |
116 | * Convert the unix version of time to dos's idea of time to be used in |
117 | * file timestamps. The passed in unix time is assumed to be in GMT. |
118 | */ |
119 | void |
120 | unix2dostime(const struct timespec *tsp, int gmtoff, u_int16_t *ddp, u_int16_t *dtp, u_int8_t *dhp) |
121 | { |
122 | u_long t; |
123 | struct clock_ymdhms ymd; |
124 | |
125 | t = tsp->tv_sec + gmtoff; /* time zone correction */ |
126 | |
127 | /* |
128 | * DOS timestamps can not represent dates before 1980. |
129 | */ |
130 | if (t < SECONDSTO1980) |
131 | goto invalid_dos_date; |
132 | |
133 | /* |
134 | * DOS granularity is 2 seconds |
135 | */ |
136 | t &= ~1; |
137 | |
138 | /* |
139 | * Convert to year/month/day/.. format |
140 | */ |
141 | clock_secs_to_ymdhms(t, &ymd); |
142 | if (ymd.dt_year > DOSMAXYEAR) |
143 | goto invalid_dos_date; |
144 | |
145 | /* |
146 | * Now transform to DOS format |
147 | */ |
148 | *ddp = (ymd.dt_day << DD_DAY_SHIFT) |
149 | + (ymd.dt_mon << DD_MONTH_SHIFT) |
150 | + ((ymd.dt_year - DOSBIASYEAR) << DD_YEAR_SHIFT); |
151 | if (dhp) |
152 | *dhp = (tsp->tv_sec & 1) * 100 + tsp->tv_nsec / 10000000; |
153 | if (dtp) |
154 | *dtp = (((t / 2) % 30) << DT_2SECONDS_SHIFT) |
155 | + (((t / 60) % 60) << DT_MINUTES_SHIFT) |
156 | + (((t / 3600) % 24) << DT_HOURS_SHIFT); |
157 | return; |
158 | |
159 | invalid_dos_date: |
160 | *ddp = 0; |
161 | if (dtp) |
162 | *dtp = 0; |
163 | if (dhp) |
164 | *dhp = 0; |
165 | } |
166 | |
167 | /* |
168 | * Convert from dos' idea of time to unix'. This will probably only be |
169 | * called from the stat(), and fstat() system calls and so probably need |
170 | * not be too efficient. |
171 | */ |
172 | void |
173 | dos2unixtime(u_int dd, u_int dt, u_int dh, int gmtoff, struct timespec *tsp) |
174 | { |
175 | time_t seconds; |
176 | struct clock_ymdhms ymd; |
177 | |
178 | if (dd == 0) { |
179 | /* |
180 | * Uninitialized field, return the epoch. |
181 | */ |
182 | tsp->tv_sec = 0; |
183 | tsp->tv_nsec = 0; |
184 | return; |
185 | } |
186 | |
187 | memset(&ymd, 0, sizeof(ymd)); |
188 | ymd.dt_year = ((dd & DD_YEAR_MASK) >> DD_YEAR_SHIFT) + 1980 ; |
189 | ymd.dt_mon = ((dd & DD_MONTH_MASK) >> DD_MONTH_SHIFT); |
190 | ymd.dt_day = ((dd & DD_DAY_MASK) >> DD_DAY_SHIFT); |
191 | ymd.dt_hour = (dt & DT_HOURS_MASK) >> DT_HOURS_SHIFT; |
192 | ymd.dt_min = (dt & DT_MINUTES_MASK) >> DT_MINUTES_SHIFT; |
193 | ymd.dt_sec = ((dt & DT_2SECONDS_MASK) >> DT_2SECONDS_SHIFT) * 2; |
194 | |
195 | seconds = clock_ymdhms_to_secs(&ymd); |
196 | |
197 | tsp->tv_sec = seconds; |
198 | tsp->tv_sec -= gmtoff; /* time zone correction */ |
199 | tsp->tv_nsec = (dh % 100) * 10000000; |
200 | } |
201 | |
202 | static const u_char |
203 | unix2dos[256] = { |
204 | 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */ |
205 | 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */ |
206 | 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */ |
207 | 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */ |
208 | 0, '!', 0, '#', '$', '%', '&', '\'', /* 20-27 */ |
209 | '(', ')', 0, '+', 0, '-', 0, 0, /* 28-2f */ |
210 | '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ |
211 | '8', '9', 0, 0, 0, 0, 0, 0, /* 38-3f */ |
212 | '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 40-47 */ |
213 | 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 48-4f */ |
214 | 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 50-57 */ |
215 | 'X', 'Y', 'Z', 0, 0, 0, '^', '_', /* 58-5f */ |
216 | '`', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 60-67 */ |
217 | 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 68-6f */ |
218 | 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 70-77 */ |
219 | 'X', 'Y', 'Z', '{', 0, '}', '~', 0, /* 78-7f */ |
220 | 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */ |
221 | 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */ |
222 | 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */ |
223 | 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */ |
224 | 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */ |
225 | 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */ |
226 | 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */ |
227 | 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */ |
228 | 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */ |
229 | 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */ |
230 | 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */ |
231 | 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */ |
232 | 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */ |
233 | 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */ |
234 | 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */ |
235 | 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */ |
236 | }; |
237 | |
238 | static const u_char |
239 | dos2unix[256] = { |
240 | '?', '?', '?', '?', '?', '?', '?', '?', /* 00-07 */ |
241 | '?', '?', '?', '?', '?', '?', '?', '?', /* 08-0f */ |
242 | '?', '?', '?', '?', '?', '?', '?', '?', /* 10-17 */ |
243 | '?', '?', '?', '?', '?', '?', '?', '?', /* 18-1f */ |
244 | ' ', '!', '"', '#', '$', '%', '&', '\'', /* 20-27 */ |
245 | '(', ')', '*', '+', ',', '-', '.', '/', /* 28-2f */ |
246 | '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ |
247 | '8', '9', ':', ';', '<', '=', '>', '?', /* 38-3f */ |
248 | '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 40-47 */ |
249 | 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 48-4f */ |
250 | 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 50-57 */ |
251 | 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', /* 58-5f */ |
252 | '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 60-67 */ |
253 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 68-6f */ |
254 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 70-77 */ |
255 | 'x', 'y', 'z', '{', '|', '}', '~', 0x7f, /* 78-7f */ |
256 | 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */ |
257 | 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */ |
258 | 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */ |
259 | 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, '?', /* 98-9f */ |
260 | 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */ |
261 | 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */ |
262 | '?', '?', '?', '?', '?', 0xc1, 0xc2, 0xc0, /* b0-b7 */ |
263 | 0xa9, '?', '?', '?', '?', 0xa2, 0xa5, '?', /* b8-bf */ |
264 | '?', '?', '?', '?', '?', '?', 0xe3, 0xc3, /* c0-c7 */ |
265 | '?', '?', '?', '?', '?', '?', '?', 0xa4, /* c8-cf */ |
266 | 0xf0, 0xd0, 0xca, 0xcb, 0xc8, '?', 0xcd, 0xce, /* d0-d7 */ |
267 | 0xcf, '?', '?', '?', '?', 0xa6, 0xcc, '?', /* d8-df */ |
268 | 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */ |
269 | 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */ |
270 | 0xad, 0xb1, '?', 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */ |
271 | 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, '?', '?', /* f8-ff */ |
272 | }; |
273 | |
274 | static const u_char |
275 | u2l[256] = { |
276 | 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ |
277 | 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ |
278 | 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ |
279 | 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ |
280 | ' ', '!', '"', '#', '$', '%', '&', '\'', /* 20-27 */ |
281 | '(', ')', '*', '+', ',', '-', '.', '/', /* 28-2f */ |
282 | '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ |
283 | '8', '9', ':', ';', '<', '=', '>', '?', /* 38-3f */ |
284 | '@', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 40-47 */ |
285 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 48-4f */ |
286 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 50-57 */ |
287 | 'x', 'y', 'z', '[', '\\', ']', '^', '_', /* 58-5f */ |
288 | '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 60-67 */ |
289 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 68-6f */ |
290 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 70-77 */ |
291 | 'x', 'y', 'z', '{', '|', '}', '~', 0x7f, /* 78-7f */ |
292 | 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ |
293 | 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ |
294 | 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ |
295 | 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ |
296 | 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ |
297 | 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ |
298 | 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ |
299 | 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ |
300 | 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ |
301 | 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ |
302 | 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ |
303 | 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ |
304 | 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ |
305 | 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ |
306 | 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ |
307 | 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ |
308 | }; |
309 | |
310 | /* |
311 | * DOS filenames are made of 2 parts, the name part and the extension part. |
312 | * The name part is 8 characters long and the extension part is 3 |
313 | * characters long. They may contain trailing blanks if the name or |
314 | * extension are not long enough to fill their respective fields. |
315 | */ |
316 | |
317 | /* |
318 | * Convert a DOS filename to a unix filename. And, return the number of |
319 | * characters in the resulting unix filename excluding the terminating |
320 | * null. |
321 | */ |
322 | int |
323 | dos2unixfn(u_char dn[11], u_char *un, int lower) |
324 | { |
325 | int i, j; |
326 | int thislong = 1; |
327 | u_char c; |
328 | |
329 | /* |
330 | * If first char of the filename is SLOT_E5 (0x05), then the real |
331 | * first char of the filename should be 0xe5. But, they couldn't |
332 | * just have a 0xe5 mean 0xe5 because that is used to mean a freed |
333 | * directory slot. Another dos quirk. |
334 | */ |
335 | if (*dn == SLOT_E5) |
336 | c = dos2unix[0xe5]; |
337 | else |
338 | c = dos2unix[*dn]; |
339 | *un++ = lower ? u2l[c] : c; |
340 | |
341 | /* |
342 | * Copy the rest into the unix filename string, ignoring |
343 | * trailing blanks. |
344 | */ |
345 | |
346 | for (j=7; (j >= 0) && (dn[j] == ' '); j--) |
347 | ; |
348 | |
349 | for (i = 1; i <= j; i++) { |
350 | c = dos2unix[dn[i]]; |
351 | *un++ = lower ? u2l[c] : c; |
352 | thislong++; |
353 | } |
354 | dn += 8; |
355 | |
356 | /* |
357 | * Now, if there is an extension then put in a period and copy in |
358 | * the extension. |
359 | */ |
360 | if (*dn != ' ') { |
361 | *un++ = '.'; |
362 | thislong++; |
363 | for (i = 0; i < 3 && *dn != ' '; i++) { |
364 | c = dos2unix[*dn++]; |
365 | *un++ = lower ? u2l[c] : c; |
366 | thislong++; |
367 | } |
368 | } |
369 | *un++ = 0; |
370 | |
371 | return (thislong); |
372 | } |
373 | |
374 | /* |
375 | * Convert a unix filename to a DOS filename according to Win95 rules. |
376 | * If applicable and gen is not 0, it is inserted into the converted |
377 | * filename as a generation number. |
378 | * Returns |
379 | * 0 if name couldn't be converted |
380 | * 1 if the converted name is the same as the original |
381 | * (no long filename entry necessary for Win95) |
382 | * 2 if conversion was successful |
383 | * 3 if conversion was successful and generation number was inserted |
384 | */ |
385 | int |
386 | unix2dosfn(const u_char *un, u_char dn[12], int unlen, u_int gen) |
387 | { |
388 | int i, j, l; |
389 | int conv = 1; |
390 | const u_char *cp, *dp, *dp1; |
391 | u_char gentext[6], *wcp; |
392 | int shortlen; |
393 | |
394 | /* |
395 | * Fill the dos filename string with blanks. These are DOS's pad |
396 | * characters. |
397 | */ |
398 | for (i = 0; i < 11; i++) |
399 | dn[i] = ' '; |
400 | dn[11] = 0; |
401 | |
402 | /* |
403 | * The filenames "." and ".." are handled specially, since they |
404 | * don't follow dos filename rules. |
405 | */ |
406 | if (un[0] == '.' && unlen == 1) { |
407 | dn[0] = '.'; |
408 | return gen <= 1; |
409 | } |
410 | if (un[0] == '.' && un[1] == '.' && unlen == 2) { |
411 | dn[0] = '.'; |
412 | dn[1] = '.'; |
413 | return gen <= 1; |
414 | } |
415 | |
416 | /* |
417 | * Filenames with only blanks and dots are not allowed! |
418 | */ |
419 | for (cp = un, i = unlen; --i >= 0; cp++) |
420 | if (*cp != ' ' && *cp != '.') |
421 | break; |
422 | if (i < 0) |
423 | return 0; |
424 | |
425 | /* |
426 | * Now find the extension |
427 | * Note: dot as first char doesn't start extension |
428 | * and trailing dots and blanks are ignored |
429 | */ |
430 | dp = dp1 = 0; |
431 | for (cp = un + 1, i = unlen - 1; --i >= 0;) { |
432 | switch (*cp++) { |
433 | case '.': |
434 | if (!dp1) |
435 | dp1 = cp; |
436 | break; |
437 | case ' ': |
438 | break; |
439 | default: |
440 | if (dp1) |
441 | dp = dp1; |
442 | dp1 = 0; |
443 | break; |
444 | } |
445 | } |
446 | |
447 | /* |
448 | * Now convert it |
449 | */ |
450 | if (dp) { |
451 | if (dp1) |
452 | l = dp1 - dp; |
453 | else |
454 | l = unlen - (dp - un); |
455 | for (i = 0, j = 8; i < l && j < 11; i++, j++) { |
456 | if (dp[i] != (dn[j] = unix2dos[dp[i]]) |
457 | && conv != 3) |
458 | conv = 2; |
459 | if (!dn[j]) { |
460 | conv = 3; |
461 | dn[j--] = ' '; |
462 | } |
463 | } |
464 | if (i < l) |
465 | conv = 3; |
466 | dp--; |
467 | } else { |
468 | for (dp = cp; *--dp == ' ' || *dp == '.';); |
469 | dp++; |
470 | } |
471 | |
472 | shortlen = (dp - un) <= 8; |
473 | |
474 | /* |
475 | * Now convert the rest of the name |
476 | */ |
477 | for (i = j = 0; un < dp && j < 8; i++, j++, un++) { |
478 | if ((*un == ' ') && shortlen) |
479 | dn[j] = ' '; |
480 | else |
481 | dn[j] = unix2dos[*un]; |
482 | if ((*un != dn[j]) |
483 | && conv != 3) |
484 | conv = 2; |
485 | if (!dn[j]) { |
486 | conv = 3; |
487 | dn[j--] = ' '; |
488 | } |
489 | } |
490 | if (un < dp) |
491 | conv = 3; |
492 | /* |
493 | * If we didn't have any chars in filename, |
494 | * generate a default |
495 | */ |
496 | if (!j) |
497 | dn[0] = '_'; |
498 | |
499 | /* |
500 | * The first character cannot be E5, |
501 | * because that means a deleted entry |
502 | */ |
503 | if (dn[0] == 0xe5) |
504 | dn[0] = SLOT_E5; |
505 | |
506 | /* |
507 | * If there wasn't any char dropped, |
508 | * there is no place for generation numbers |
509 | */ |
510 | if (conv != 3) { |
511 | if (gen > 1) |
512 | return 0; |
513 | return conv; |
514 | } |
515 | |
516 | /* |
517 | * Now insert the generation number into the filename part |
518 | */ |
519 | for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) |
520 | *--wcp = gen % 10 + '0'; |
521 | if (gen) |
522 | return 0; |
523 | for (i = 8; dn[--i] == ' ';); |
524 | i++; |
525 | if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) |
526 | i = 8 - (gentext + sizeof(gentext) - wcp + 1); |
527 | dn[i++] = '~'; |
528 | while (wcp < gentext + sizeof(gentext)) |
529 | dn[i++] = *wcp++; |
530 | return 3; |
531 | } |
532 | |
533 | /* |
534 | * Create a Win95 long name directory entry |
535 | * Note: assumes that the filename is valid, |
536 | * i.e. doesn't consist solely of blanks and dots |
537 | */ |
538 | int |
539 | unix2winfn(const u_char *un, int unlen, struct winentry *wep, int cnt, int chksum, int utf8) |
540 | { |
541 | u_int16_t wn[WIN_MAXLEN], *p; |
542 | int i, len; |
543 | const u_char *cp; |
544 | |
545 | /* |
546 | * Drop trailing blanks and dots |
547 | */ |
548 | for (cp = un + unlen; unlen > 0; unlen--) |
549 | if (*--cp != ' ' && *cp != '.') |
550 | break; |
551 | |
552 | /* |
553 | * Offset of this entry |
554 | */ |
555 | i = (cnt - 1) * WIN_CHARS; |
556 | |
557 | /* |
558 | * Translate UNIX name to ucs-2 |
559 | */ |
560 | len = utf8 ? utf8ucs2str(un, unlen, wn, WIN_MAXLEN) : char8ucs2str(un, unlen, wn, WIN_MAXLEN); |
561 | ucs2pad(wn, len, WIN_MAXLEN); |
562 | |
563 | /* |
564 | * Initialize winentry to some useful default |
565 | */ |
566 | memset(wep, 0xff, sizeof(*wep)); |
567 | wep->weCnt = cnt; |
568 | wep->weAttributes = ATTR_WIN95; |
569 | wep->weReserved1 = 0; |
570 | wep->weChksum = chksum; |
571 | wep->weReserved2 = 0; |
572 | |
573 | /* |
574 | * Store name segment into directory entry |
575 | */ |
576 | p = &wn[i]; |
577 | memcpy(wep->wePart1, p, sizeof(wep->wePart1)); |
578 | p += sizeof(wep->wePart1) / sizeof(*p); |
579 | memcpy(wep->wePart2, p, sizeof(wep->wePart2)); |
580 | p += sizeof(wep->wePart2) / sizeof(*p); |
581 | memcpy(wep->wePart3, p, sizeof(wep->wePart3)); |
582 | |
583 | if (len > i + WIN_CHARS) |
584 | return 1; |
585 | |
586 | wep->weCnt |= WIN_LAST; |
587 | return 0; |
588 | } |
589 | |
590 | /* |
591 | * Compare our filename to the one in the Win95 entry |
592 | * Returns the checksum or -1 if no match |
593 | */ |
594 | int |
595 | winChkName(const u_char *un, int unlen, struct winentry *wep, int chksum, int utf8) |
596 | { |
597 | u_int16_t wn[WIN_MAXLEN], *p; |
598 | u_int16_t buf[WIN_CHARS]; |
599 | int i, len; |
600 | |
601 | /* |
602 | * First compare checksums |
603 | */ |
604 | if (wep->weCnt & WIN_LAST) |
605 | chksum = wep->weChksum; |
606 | else if (chksum != wep->weChksum) |
607 | chksum = -1; |
608 | if (chksum == -1) |
609 | return -1; |
610 | |
611 | /* |
612 | * Offset of this entry |
613 | */ |
614 | i = ((wep->weCnt & WIN_CNT) - 1) * WIN_CHARS; |
615 | |
616 | /* |
617 | * Translate UNIX name to ucs-2 |
618 | */ |
619 | len = utf8 ? utf8ucs2str(un, unlen, wn, WIN_MAXLEN) : char8ucs2str(un, unlen, wn, WIN_MAXLEN); |
620 | ucs2pad(wn, len, WIN_MAXLEN); |
621 | |
622 | if (i >= len + 1) |
623 | return -1; |
624 | if ((wep->weCnt & WIN_LAST) && (len - i > WIN_CHARS)) |
625 | return -1; |
626 | |
627 | /* |
628 | * Fetch name segment from directory entry |
629 | */ |
630 | p = &buf[0]; |
631 | memcpy(p, wep->wePart1, sizeof(wep->wePart1)); |
632 | p += sizeof(wep->wePart1) / sizeof(*p); |
633 | memcpy(p, wep->wePart2, sizeof(wep->wePart2)); |
634 | p += sizeof(wep->wePart2) / sizeof(*p); |
635 | memcpy(p, wep->wePart3, sizeof(wep->wePart3)); |
636 | |
637 | /* |
638 | * And compare name segment |
639 | */ |
640 | if (! (utf8 ? ucs2match(&wn[i], buf, WIN_CHARS) : char8match(&wn[i], buf, WIN_CHARS))) |
641 | return -1; |
642 | |
643 | return chksum; |
644 | } |
645 | |
646 | /* |
647 | * Convert Win95 filename to dirbuf. |
648 | * Returns the checksum or -1 if impossible |
649 | */ |
650 | int |
651 | win2unixfn(struct winentry *wep, struct dirent *dp, int chksum, |
652 | uint16_t *namlen, int utf8) |
653 | { |
654 | u_int16_t wn[WIN_CHARS], *p; |
655 | u_int8_t buf[WIN_CHARS*3]; |
656 | int len; |
657 | |
658 | if ((wep->weCnt & WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) |
659 | || !(wep->weCnt & WIN_CNT)) |
660 | return -1; |
661 | |
662 | /* |
663 | * First compare checksums |
664 | */ |
665 | if (wep->weCnt & WIN_LAST) { |
666 | chksum = wep->weChksum; |
667 | *namlen = 0; |
668 | } else if (chksum != wep->weChksum) |
669 | chksum = -1; |
670 | if (chksum == -1) |
671 | return -1; |
672 | |
673 | /* |
674 | * Fetch name segment from directory entry |
675 | */ |
676 | p = &wn[0]; |
677 | memcpy(p, wep->wePart1, sizeof(wep->wePart1)); |
678 | p += sizeof(wep->wePart1) / sizeof(*p); |
679 | memcpy(p, wep->wePart2, sizeof(wep->wePart2)); |
680 | p += sizeof(wep->wePart2) / sizeof(*p); |
681 | memcpy(p, wep->wePart3, sizeof(wep->wePart3)); |
682 | |
683 | /* |
684 | * Don't allow slashes in UNIX names. Discard that entry. |
685 | */ |
686 | if (invalidname(wn, WIN_CHARS)) |
687 | return -1; |
688 | |
689 | /* |
690 | * Translate ucs-2 to UNIX name |
691 | */ |
692 | len = utf8 ? ucs2utf8str(wn, WIN_CHARS, buf, sizeof(buf)) |
693 | : ucs2char8str(wn, WIN_CHARS, buf, sizeof(buf)); |
694 | |
695 | KASSERT(len >= 0); |
696 | KASSERT((size_t)len <= MIN(sizeof(buf), sizeof(dp->d_name)-1)); |
697 | |
698 | /* |
699 | * Prepend name segment to directory entry |
700 | * |
701 | * This ignores the slot number from the windows entry but |
702 | * assumes that segments are read in reverse order. |
703 | * |
704 | * The UCS-2 name (up to 255 chars) can overflow the UNIX |
705 | * directory entry (up to 511 bytes). Trailing characters |
706 | * are silently discarded. This could also end in multiple |
707 | * files using the same (truncated) name. |
708 | */ |
709 | *namlen += len; |
710 | if (*namlen > sizeof(dp->d_name) - 1) |
711 | *namlen = sizeof(dp->d_name) - 1; |
712 | |
713 | KASSERT(*namlen >= len); |
714 | |
715 | memmove(&dp->d_name[len], &dp->d_name[0], *namlen - len); |
716 | memcpy(dp->d_name, buf, len); |
717 | |
718 | return chksum; |
719 | } |
720 | |
721 | /* |
722 | * Compute the checksum of a DOS filename for Win95 use |
723 | */ |
724 | u_int8_t |
725 | winChksum(u_int8_t *name) |
726 | { |
727 | int i; |
728 | u_int8_t s; |
729 | |
730 | for (s = 0, i = 11; --i >= 0; s += *name++) |
731 | s = (s << 7) | (s >> 1); |
732 | return s; |
733 | } |
734 | |
735 | /* |
736 | * Determine the number of slots necessary for Win95 names |
737 | */ |
738 | int |
739 | winSlotCnt(const u_char *un, int unlen, int utf8) |
740 | { |
741 | const u_char *cp; |
742 | int len; |
743 | |
744 | /* |
745 | * Drop trailing blanks and dots |
746 | */ |
747 | for (cp = un + unlen; unlen > 0; unlen--) |
748 | if (*--cp != ' ' && *cp != '.') |
749 | break; |
750 | |
751 | len = utf8 ? utf8ucs2str(un, unlen, NULL, WIN_MAXLEN) : unlen; |
752 | |
753 | return howmany(len, WIN_CHARS); |
754 | } |
755 | |
756 | /* |
757 | * Scan windows name for characters that must not |
758 | * appear in a UNIX filename |
759 | */ |
760 | static int |
761 | invalidname(const u_int16_t *in, int n) |
762 | { |
763 | while (n-- > 0) { |
764 | if (*in++ == '/') |
765 | return 1; |
766 | } |
767 | |
768 | return 0; |
769 | } |
770 | |
771 | /* |
772 | * Convert UCS-2 character into UTF-8 |
773 | * return number of output bytes or 0 if output |
774 | * buffer is too short |
775 | */ |
776 | static int |
777 | ucs2utf8(const u_int16_t *in, u_int8_t *out, int n) |
778 | { |
779 | uint16_t inch = le16toh(in[0]); |
780 | |
781 | if (inch <= 0x007f) { |
782 | if (n < 1) return 0; |
783 | if (out) |
784 | *out++ = inch; |
785 | return 1; |
786 | } else if (inch <= 0x07ff) { |
787 | if (n < 2) return 0; |
788 | if (out) { |
789 | *out++ = 0xc0 | (inch >> 6); |
790 | *out++ = 0x80 | (inch & 0x3f); |
791 | } |
792 | return 2; |
793 | } else { |
794 | if (n < 3) return 0; |
795 | if (out) { |
796 | *out++ = 0xe0 | (inch >> 12); |
797 | *out++ = 0x80 | ((inch >> 6) & 0x3f); |
798 | *out++ = 0x80 | (inch & 0x3f); |
799 | } |
800 | return 3; |
801 | } |
802 | } |
803 | |
804 | |
805 | /* |
806 | * Convert UTF-8 bytes into UCS-2 character |
807 | * return number of input bytes, 0 if input |
808 | * is too short and -1 if input is invalid |
809 | */ |
810 | static int |
811 | utf8ucs2(const u_int8_t *in, int n, u_int16_t *out) |
812 | { |
813 | uint16_t outch; |
814 | |
815 | if (n < 1) return 0; |
816 | |
817 | if (in[0] <= 0x7f) { |
818 | outch = in[0]; |
819 | if (out) |
820 | *out = htole16(outch); |
821 | return 1; |
822 | } else if (in[0] <= 0xdf) { |
823 | if (n < 2) return 0; |
824 | outch = (in[0] & 0x1f) << 6 | (in[1] & 0x3f); |
825 | if (out) |
826 | *out = htole16(outch); |
827 | return 2; |
828 | } else if (in[0] <= 0xef) { |
829 | if (n < 3) return 0; |
830 | outch = (in[0] & 0x1f) << 12 | (in[1] & 0x3f) << 6 | (in[2] & 0x3f); |
831 | if (out) |
832 | *out = htole16(outch); |
833 | return 3; |
834 | } |
835 | |
836 | return -1; |
837 | } |
838 | |
839 | /* |
840 | * Convert UCS-2 string into UTF-8 string |
841 | * return total number of output bytes |
842 | */ |
843 | static int |
844 | ucs2utf8str(const u_int16_t *in, int n, u_int8_t *out, int m) |
845 | { |
846 | u_int8_t *p; |
847 | int outlen; |
848 | |
849 | p = out; |
850 | while (n > 0 && *in != 0) { |
851 | outlen = ucs2utf8(in, out ? p : out, m); |
852 | if (outlen == 0) |
853 | break; |
854 | p += outlen; |
855 | m -= outlen; |
856 | in += 1; |
857 | n -= 1; |
858 | } |
859 | |
860 | return p - out; |
861 | } |
862 | |
863 | /* |
864 | * Convert UTF8 string into UCS-2 string |
865 | * return total number of output chacters |
866 | */ |
867 | static int |
868 | utf8ucs2str(const u_int8_t *in, int n, u_int16_t *out, int m) |
869 | { |
870 | u_int16_t *p; |
871 | int inlen; |
872 | |
873 | p = out; |
874 | while (n > 0 && *in != 0) { |
875 | if (m < 1) |
876 | break; |
877 | inlen = utf8ucs2(in, n, out ? p : out); |
878 | if (inlen <= 0) |
879 | break; |
880 | in += inlen; |
881 | n -= inlen; |
882 | p += 1; |
883 | m -= 1; |
884 | } |
885 | |
886 | return p - out; |
887 | } |
888 | |
889 | /* |
890 | * Convert UCS-2 string into 8bit character string |
891 | * return total number of output bytes |
892 | */ |
893 | static int |
894 | ucs2char8str(const u_int16_t *in, int n, u_int8_t *out, int m) |
895 | { |
896 | u_int8_t *p; |
897 | u_int16_t inch; |
898 | |
899 | p = out; |
900 | while (n > 0 && in[0] != 0) { |
901 | if (m < 1) |
902 | break; |
903 | inch = le16toh(in[0]); |
904 | if (inch > 255) |
905 | break; |
906 | if (p) |
907 | p[0] = inch; |
908 | p += 1; |
909 | m -= 1; |
910 | in += 1; |
911 | n -= 1; |
912 | } |
913 | |
914 | return p - out; |
915 | } |
916 | |
917 | /* |
918 | * Convert 8bit character string into UCS-2 string |
919 | * return total number of output chacters |
920 | */ |
921 | static int |
922 | char8ucs2str(const u_int8_t *in, int n, u_int16_t *out, int m) |
923 | { |
924 | u_int16_t *p; |
925 | |
926 | p = out; |
927 | while (n > 0 && in[0] != 0) { |
928 | if (m < 1) |
929 | break; |
930 | if (p) |
931 | p[0] = htole16(in[0]); |
932 | p += 1; |
933 | m -= 1; |
934 | in += 1; |
935 | n -= 1; |
936 | } |
937 | |
938 | return p - out; |
939 | } |
940 | |
941 | static void |
942 | ucs2pad(u_int16_t *buf, int len, int size) |
943 | { |
944 | |
945 | if (len < size-1) |
946 | buf[len++] = 0x0000; |
947 | while (len < size) |
948 | buf[len++] = 0xffff; |
949 | } |
950 | |
951 | /* |
952 | * Fold UCS-2 character to uppercase |
953 | */ |
954 | static u_int16_t |
955 | ucs2fold(u_int16_t w) |
956 | { |
957 | int low,high,mid; |
958 | u_int16_t check; |
959 | extern const u_int16_t msdosfs_unicode_foldmap[]; |
960 | extern size_t msdosfs_unicode_foldmap_entries; |
961 | |
962 | w = le16toh(w); |
963 | |
964 | low = 0; |
965 | high = msdosfs_unicode_foldmap_entries / 2; |
966 | while (low < high) { |
967 | mid = (low + high)/2; |
968 | check = msdosfs_unicode_foldmap[2*mid+0]; |
969 | |
970 | if (w == check) { |
971 | w = msdosfs_unicode_foldmap[2*mid+1]; |
972 | break; |
973 | } |
974 | |
975 | if (w < check) |
976 | high = mid; |
977 | else |
978 | low = mid+1; |
979 | } |
980 | |
981 | w = le16toh(w); |
982 | |
983 | return w; |
984 | } |
985 | |
986 | /* |
987 | * Compare two UCS-2 strings case-insensitive |
988 | * |
989 | * uses the Unicode case folding table |
990 | */ |
991 | static int |
992 | ucs2match(u_int16_t *w1, u_int16_t *w2, int n) |
993 | { |
994 | u_int16_t u1, u2; |
995 | |
996 | while (n > 0) { |
997 | if (*w1 == 0 || *w2 == 0) |
998 | return *w1 == *w2; |
999 | u1 = ucs2fold(*w1); |
1000 | u2 = ucs2fold(*w2); |
1001 | if (u1 != u2) |
1002 | return 0; |
1003 | ++w1; |
1004 | ++w2; |
1005 | --n; |
1006 | } |
1007 | |
1008 | return 1; |
1009 | } |
1010 | |
1011 | /* |
1012 | * Compare two 8bit char conversions case-insensitive |
1013 | * |
1014 | * uses the DOS case folding table |
1015 | */ |
1016 | static int |
1017 | char8match(u_int16_t *w1, u_int16_t *w2, int n) |
1018 | { |
1019 | u_int16_t u1, u2; |
1020 | |
1021 | while (n > 0) { |
1022 | u1 = le16toh(*w1); |
1023 | u2 = le16toh(*w2); |
1024 | if (u1 == 0 || u2 == 0) |
1025 | return u1 == u2; |
1026 | if (u1 > 255 || u2 > 255) |
1027 | return 0; |
1028 | u1 = u2l[u1 & 0xff]; |
1029 | u2 = u2l[u2 & 0xff]; |
1030 | if (u1 != u2) |
1031 | return 0; |
1032 | ++w1; |
1033 | ++w2; |
1034 | --n; |
1035 | } |
1036 | |
1037 | return 1; |
1038 | } |
1039 | |
1040 | |