1 /* $NetBSD: msdosfs_conv.c,v 1.20 2023/06/02 08:51:47 andvar 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 1.20 2023/06/02 08:51:47 andvar 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 beyond 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 msdosfs_unix2dostime(const struct timespec *tsp, int gmtoff, 121 u_int16_t *ddp, u_int16_t *dtp, u_int8_t *dhp) 122 { 123 u_long t; 124 struct clock_ymdhms ymd; 125 126 t = tsp->tv_sec + gmtoff; /* time zone correction */ 127 128 /* 129 * DOS timestamps can not represent dates before 1980. 130 */ 131 if (t < SECONDSTO1980) 132 goto invalid_dos_date; 133 134 /* 135 * DOS granularity is 2 seconds 136 */ 137 t &= ~1; 138 139 /* 140 * Convert to year/month/day/.. format 141 */ 142 clock_secs_to_ymdhms(t, &ymd); 143 if (ymd.dt_year > DOSMAXYEAR) 144 goto invalid_dos_date; 145 146 /* 147 * Now transform to DOS format 148 */ 149 *ddp = (ymd.dt_day << DD_DAY_SHIFT) 150 + (ymd.dt_mon << DD_MONTH_SHIFT) 151 + ((ymd.dt_year - DOSBIASYEAR) << DD_YEAR_SHIFT); 152 if (dhp) 153 *dhp = (tsp->tv_sec & 1) * 100 + tsp->tv_nsec / 10000000; 154 if (dtp) 155 *dtp = (((t / 2) % 30) << DT_2SECONDS_SHIFT) 156 + (((t / 60) % 60) << DT_MINUTES_SHIFT) 157 + (((t / 3600) % 24) << DT_HOURS_SHIFT); 158 return; 159 160 invalid_dos_date: 161 *ddp = 0; 162 if (dtp) 163 *dtp = 0; 164 if (dhp) 165 *dhp = 0; 166 } 167 168 /* 169 * Convert from dos' idea of time to unix'. This will probably only be 170 * called from the stat(), and fstat() system calls and so probably need 171 * not be too efficient. 172 */ 173 void 174 msdosfs_dos2unixtime(u_int dd, u_int dt, u_int dh, int gmtoff, 175 struct timespec *tsp) 176 { 177 time_t seconds; 178 struct clock_ymdhms ymd; 179 180 if (dd == 0) { 181 /* 182 * Uninitialized field, return the epoch. 183 */ 184 tsp->tv_sec = 0; 185 tsp->tv_nsec = 0; 186 return; 187 } 188 189 memset(&ymd, 0, sizeof(ymd)); 190 ymd.dt_year = ((dd & DD_YEAR_MASK) >> DD_YEAR_SHIFT) + 1980 ; 191 ymd.dt_mon = ((dd & DD_MONTH_MASK) >> DD_MONTH_SHIFT); 192 ymd.dt_day = ((dd & DD_DAY_MASK) >> DD_DAY_SHIFT); 193 ymd.dt_hour = (dt & DT_HOURS_MASK) >> DT_HOURS_SHIFT; 194 ymd.dt_min = (dt & DT_MINUTES_MASK) >> DT_MINUTES_SHIFT; 195 ymd.dt_sec = ((dt & DT_2SECONDS_MASK) >> DT_2SECONDS_SHIFT) * 2; 196 197 seconds = clock_ymdhms_to_secs(&ymd); 198 199 tsp->tv_sec = seconds; 200 tsp->tv_sec -= gmtoff; /* time zone correction */ 201 tsp->tv_nsec = (dh % 100) * 10000000; 202 } 203 204 static const u_char 205 unix2dos[256] = { 206 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */ 207 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */ 208 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */ 209 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */ 210 0, '!', 0, '#', '$', '%', '&', '\'', /* 20-27 */ 211 '(', ')', 0, '+', 0, '-', 0, 0, /* 28-2f */ 212 '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ 213 '8', '9', 0, 0, 0, 0, 0, 0, /* 38-3f */ 214 '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 40-47 */ 215 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 48-4f */ 216 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 50-57 */ 217 'X', 'Y', 'Z', 0, 0, 0, '^', '_', /* 58-5f */ 218 '`', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 60-67 */ 219 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 68-6f */ 220 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 70-77 */ 221 'X', 'Y', 'Z', '{', 0, '}', '~', 0, /* 78-7f */ 222 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */ 223 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */ 224 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */ 225 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */ 226 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */ 227 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */ 228 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */ 229 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */ 230 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */ 231 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */ 232 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */ 233 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */ 234 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */ 235 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */ 236 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */ 237 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */ 238 }; 239 240 static const u_char 241 dos2unix[256] = { 242 '?', '?', '?', '?', '?', '?', '?', '?', /* 00-07 */ 243 '?', '?', '?', '?', '?', '?', '?', '?', /* 08-0f */ 244 '?', '?', '?', '?', '?', '?', '?', '?', /* 10-17 */ 245 '?', '?', '?', '?', '?', '?', '?', '?', /* 18-1f */ 246 ' ', '!', '"', '#', '$', '%', '&', '\'', /* 20-27 */ 247 '(', ')', '*', '+', ',', '-', '.', '/', /* 28-2f */ 248 '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ 249 '8', '9', ':', ';', '<', '=', '>', '?', /* 38-3f */ 250 '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', /* 40-47 */ 251 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', /* 48-4f */ 252 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', /* 50-57 */ 253 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', /* 58-5f */ 254 '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 60-67 */ 255 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 68-6f */ 256 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 70-77 */ 257 'x', 'y', 'z', '{', '|', '}', '~', 0x7f, /* 78-7f */ 258 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */ 259 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */ 260 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */ 261 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, '?', /* 98-9f */ 262 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */ 263 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */ 264 '?', '?', '?', '?', '?', 0xc1, 0xc2, 0xc0, /* b0-b7 */ 265 0xa9, '?', '?', '?', '?', 0xa2, 0xa5, '?', /* b8-bf */ 266 '?', '?', '?', '?', '?', '?', 0xe3, 0xc3, /* c0-c7 */ 267 '?', '?', '?', '?', '?', '?', '?', 0xa4, /* c8-cf */ 268 0xf0, 0xd0, 0xca, 0xcb, 0xc8, '?', 0xcd, 0xce, /* d0-d7 */ 269 0xcf, '?', '?', '?', '?', 0xa6, 0xcc, '?', /* d8-df */ 270 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */ 271 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */ 272 0xad, 0xb1, '?', 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */ 273 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, '?', '?', /* f8-ff */ 274 }; 275 276 static const u_char 277 u2l[256] = { 278 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 279 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 280 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 281 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 282 ' ', '!', '"', '#', '$', '%', '&', '\'', /* 20-27 */ 283 '(', ')', '*', '+', ',', '-', '.', '/', /* 28-2f */ 284 '0', '1', '2', '3', '4', '5', '6', '7', /* 30-37 */ 285 '8', '9', ':', ';', '<', '=', '>', '?', /* 38-3f */ 286 '@', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 40-47 */ 287 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 48-4f */ 288 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 50-57 */ 289 'x', 'y', 'z', '[', '\\', ']', '^', '_', /* 58-5f */ 290 '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', /* 60-67 */ 291 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', /* 68-6f */ 292 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', /* 70-77 */ 293 'x', 'y', 'z', '{', '|', '}', '~', 0x7f, /* 78-7f */ 294 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 295 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 296 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 297 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 298 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 299 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 300 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 301 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 302 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 303 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 304 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 305 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 306 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 307 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 308 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 309 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 310 }; 311 312 /* 313 * DOS filenames are made of 2 parts, the name part and the extension part. 314 * The name part is 8 characters long and the extension part is 3 315 * characters long. They may contain trailing blanks if the name or 316 * extension are not long enough to fill their respective fields. 317 */ 318 319 /* 320 * Convert a DOS filename to a unix filename. And, return the number of 321 * characters in the resulting unix filename excluding the terminating 322 * null. 323 */ 324 int 325 msdosfs_dos2unixfn(u_char dn[11], u_char *un, int lower) 326 { 327 int i, j; 328 int thislong = 1; 329 u_char c; 330 331 /* 332 * If first char of the filename is SLOT_E5 (0x05), then the real 333 * first char of the filename should be 0xe5. But, they couldn't 334 * just have a 0xe5 mean 0xe5 because that is used to mean a freed 335 * directory slot. Another dos quirk. 336 */ 337 if (*dn == SLOT_E5) 338 c = dos2unix[0xe5]; 339 else 340 c = dos2unix[*dn]; 341 *un++ = lower ? u2l[c] : c; 342 343 /* 344 * Copy the rest into the unix filename string, ignoring 345 * trailing blanks. 346 */ 347 348 for (j=7; (j >= 0) && (dn[j] == ' '); j--) 349 ; 350 351 for (i = 1; i <= j; i++) { 352 c = dos2unix[dn[i]]; 353 *un++ = lower ? u2l[c] : c; 354 thislong++; 355 } 356 dn += 8; 357 358 /* 359 * Now, if there is an extension then put in a period and copy in 360 * the extension. 361 */ 362 if (*dn != ' ') { 363 *un++ = '.'; 364 thislong++; 365 for (i = 0; i < 3 && *dn != ' '; i++) { 366 c = dos2unix[*dn++]; 367 *un++ = lower ? u2l[c] : c; 368 thislong++; 369 } 370 } 371 *un++ = 0; 372 373 return (thislong); 374 } 375 376 /* 377 * Convert a unix filename to a DOS filename according to Win95 rules. 378 * If applicable and gen is not 0, it is inserted into the converted 379 * filename as a generation number. 380 * Returns 381 * 0 if name couldn't be converted 382 * 1 if the converted name is the same as the original 383 * (no long filename entry necessary for Win95) 384 * 2 if conversion was successful 385 * 3 if conversion was successful and generation number was inserted 386 */ 387 int 388 msdosfs_unix2dosfn(const u_char *un, u_char dn[12], int unlen, u_int gen) 389 { 390 int i, j, l; 391 int conv = 1; 392 const u_char *cp, *dp, *dp1; 393 u_char gentext[6], *wcp; 394 int shortlen; 395 396 /* 397 * Fill the dos filename string with blanks. These are DOS's pad 398 * characters. 399 */ 400 for (i = 0; i < 11; i++) 401 dn[i] = ' '; 402 dn[11] = 0; 403 404 /* 405 * The filenames "." and ".." are handled specially, since they 406 * don't follow dos filename rules. 407 */ 408 if (un[0] == '.' && unlen == 1) { 409 dn[0] = '.'; 410 return gen <= 1; 411 } 412 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 413 dn[0] = '.'; 414 dn[1] = '.'; 415 return gen <= 1; 416 } 417 418 /* 419 * Filenames with only blanks and dots are not allowed! 420 */ 421 for (cp = un, i = unlen; --i >= 0; cp++) 422 if (*cp != ' ' && *cp != '.') 423 break; 424 if (i < 0) 425 return 0; 426 427 /* 428 * Now find the extension 429 * Note: dot as first char doesn't start extension 430 * and trailing dots and blanks are ignored 431 */ 432 dp = dp1 = 0; 433 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 434 switch (*cp++) { 435 case '.': 436 if (!dp1) 437 dp1 = cp; 438 break; 439 case ' ': 440 break; 441 default: 442 if (dp1) 443 dp = dp1; 444 dp1 = 0; 445 break; 446 } 447 } 448 449 /* 450 * Now convert it 451 */ 452 if (dp) { 453 if (dp1) 454 l = dp1 - dp; 455 else 456 l = unlen - (dp - un); 457 for (i = 0, j = 8; i < l && j < 11; i++, j++) { 458 if (dp[i] != (dn[j] = unix2dos[dp[i]]) 459 && conv != 3) 460 conv = 2; 461 if (!dn[j]) { 462 conv = 3; 463 dn[j--] = ' '; 464 } 465 } 466 if (i < l) 467 conv = 3; 468 dp--; 469 } else { 470 for (dp = cp; *--dp == ' ' || *dp == '.';); 471 dp++; 472 } 473 474 shortlen = (dp - un) <= 8; 475 476 /* 477 * Now convert the rest of the name 478 */ 479 for (i = j = 0; un < dp && j < 8; i++, j++, un++) { 480 if ((*un == ' ') && shortlen) 481 dn[j] = ' '; 482 else 483 dn[j] = unix2dos[*un]; 484 if ((*un != dn[j]) 485 && conv != 3) 486 conv = 2; 487 if (!dn[j]) { 488 conv = 3; 489 dn[j--] = ' '; 490 } 491 } 492 if (un < dp) 493 conv = 3; 494 /* 495 * If we didn't have any chars in filename, 496 * generate a default 497 */ 498 if (!j) 499 dn[0] = '_'; 500 501 /* 502 * The first character cannot be E5, 503 * because that means a deleted entry 504 */ 505 if (dn[0] == 0xe5) 506 dn[0] = SLOT_E5; 507 508 /* 509 * If there wasn't any char dropped, 510 * there is no place for generation numbers 511 */ 512 if (conv != 3) { 513 if (gen > 1) 514 return 0; 515 return conv; 516 } 517 518 /* 519 * Now insert the generation number into the filename part 520 */ 521 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 522 *--wcp = gen % 10 + '0'; 523 if (gen) 524 return 0; 525 for (i = 8; dn[--i] == ' ';); 526 i++; 527 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 528 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 529 dn[i++] = '~'; 530 while (wcp < gentext + sizeof(gentext)) 531 dn[i++] = *wcp++; 532 return 3; 533 } 534 535 /* 536 * Create a Win95 long name directory entry 537 * Note: assumes that the filename is valid, 538 * i.e. doesn't consist solely of blanks and dots 539 */ 540 int 541 msdosfs_unix2winfn(const u_char *un, int unlen, struct winentry *wep, int cnt, 542 int chksum, int utf8) 543 { 544 u_int16_t wn[WIN_MAXLEN], *p; 545 int i, len; 546 const u_char *cp; 547 548 /* 549 * Drop trailing blanks and dots 550 */ 551 for (cp = un + unlen; unlen > 0; unlen--) 552 if (*--cp != ' ' && *cp != '.') 553 break; 554 555 /* 556 * Offset of this entry 557 */ 558 i = (cnt - 1) * WIN_CHARS; 559 560 /* 561 * Translate UNIX name to ucs-2 562 */ 563 len = utf8 ? utf8ucs2str(un, unlen, wn, WIN_MAXLEN) : char8ucs2str(un, unlen, wn, WIN_MAXLEN); 564 ucs2pad(wn, len, WIN_MAXLEN); 565 566 /* 567 * Initialize winentry to some useful default 568 */ 569 memset(wep, 0xff, sizeof(*wep)); 570 wep->weCnt = cnt; 571 wep->weAttributes = ATTR_WIN95; 572 wep->weReserved1 = 0; 573 wep->weChksum = chksum; 574 wep->weReserved2 = 0; 575 576 /* 577 * Store name segment into directory entry 578 */ 579 p = &wn[i]; 580 memcpy(wep->wePart1, p, sizeof(wep->wePart1)); 581 p += sizeof(wep->wePart1) / sizeof(*p); 582 memcpy(wep->wePart2, p, sizeof(wep->wePart2)); 583 p += sizeof(wep->wePart2) / sizeof(*p); 584 memcpy(wep->wePart3, p, sizeof(wep->wePart3)); 585 586 if (len > i + WIN_CHARS) 587 return 1; 588 589 wep->weCnt |= WIN_LAST; 590 return 0; 591 } 592 593 /* 594 * Compare our filename to the one in the Win95 entry 595 * Returns the checksum or -1 if no match 596 */ 597 int 598 msdosfs_winChkName(const u_char *un, int unlen, struct winentry *wep, 599 int chksum, int utf8) 600 { 601 u_int16_t wn[WIN_MAXLEN], *p; 602 u_int16_t buf[WIN_CHARS]; 603 int i, len; 604 605 /* 606 * First compare checksums 607 */ 608 if (wep->weCnt & WIN_LAST) 609 chksum = wep->weChksum; 610 else if (chksum != wep->weChksum) 611 chksum = -1; 612 if (chksum == -1) 613 return -1; 614 615 /* 616 * Offset of this entry 617 */ 618 i = ((wep->weCnt & WIN_CNT) - 1) * WIN_CHARS; 619 620 /* 621 * Translate UNIX name to ucs-2 622 */ 623 len = utf8 ? utf8ucs2str(un, unlen, wn, WIN_MAXLEN) : char8ucs2str(un, unlen, wn, WIN_MAXLEN); 624 ucs2pad(wn, len, WIN_MAXLEN); 625 626 if (i >= len + 1) 627 return -1; 628 if ((wep->weCnt & WIN_LAST) && (len - i > WIN_CHARS)) 629 return -1; 630 631 /* 632 * Fetch name segment from directory entry 633 */ 634 p = &buf[0]; 635 memcpy(p, wep->wePart1, sizeof(wep->wePart1)); 636 p += sizeof(wep->wePart1) / sizeof(*p); 637 memcpy(p, wep->wePart2, sizeof(wep->wePart2)); 638 p += sizeof(wep->wePart2) / sizeof(*p); 639 memcpy(p, wep->wePart3, sizeof(wep->wePart3)); 640 641 /* 642 * And compare name segment 643 */ 644 if (! (utf8 ? ucs2match(&wn[i], buf, WIN_CHARS) : char8match(&wn[i], buf, WIN_CHARS))) 645 return -1; 646 647 return chksum; 648 } 649 650 /* 651 * Convert Win95 filename to dirbuf. 652 * Returns the checksum or -1 if impossible 653 */ 654 int 655 msdosfs_win2unixfn(struct winentry *wep, struct dirent *dp, int chksum, 656 uint16_t *namlen, int utf8) 657 { 658 u_int16_t wn[WIN_CHARS], *p; 659 u_int8_t buf[WIN_CHARS*3]; 660 int len; 661 662 if ((wep->weCnt & WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 663 || !(wep->weCnt & WIN_CNT)) 664 return -1; 665 666 /* 667 * First compare checksums 668 */ 669 if (wep->weCnt & WIN_LAST) { 670 chksum = wep->weChksum; 671 *namlen = 0; 672 } else if (chksum != wep->weChksum) 673 chksum = -1; 674 if (chksum == -1) 675 return -1; 676 677 /* 678 * Fetch name segment from directory entry 679 */ 680 p = &wn[0]; 681 memcpy(p, wep->wePart1, sizeof(wep->wePart1)); 682 p += sizeof(wep->wePart1) / sizeof(*p); 683 memcpy(p, wep->wePart2, sizeof(wep->wePart2)); 684 p += sizeof(wep->wePart2) / sizeof(*p); 685 memcpy(p, wep->wePart3, sizeof(wep->wePart3)); 686 687 /* 688 * Don't allow slashes in UNIX names. Discard that entry. 689 */ 690 if (invalidname(wn, WIN_CHARS)) 691 return -1; 692 693 /* 694 * Translate ucs-2 to UNIX name 695 */ 696 len = utf8 ? ucs2utf8str(wn, WIN_CHARS, buf, sizeof(buf)) 697 : ucs2char8str(wn, WIN_CHARS, buf, sizeof(buf)); 698 699 KASSERT(len >= 0); 700 KASSERT((size_t)len <= MIN(sizeof(buf), sizeof(dp->d_name)-1)); 701 702 /* 703 * Prepend name segment to directory entry 704 * 705 * This ignores the slot number from the windows entry but 706 * assumes that segments are read in reverse order. 707 * 708 * The UCS-2 name (up to 255 chars) can overflow the UNIX 709 * directory entry (up to 511 bytes). Trailing characters 710 * are silently discarded. This could also end in multiple 711 * files using the same (truncated) name. 712 */ 713 *namlen += len; 714 if (*namlen > sizeof(dp->d_name) - 1) 715 *namlen = sizeof(dp->d_name) - 1; 716 717 KASSERT(*namlen >= len); 718 719 memmove(&dp->d_name[len], &dp->d_name[0], *namlen - len); 720 memcpy(dp->d_name, buf, len); 721 722 return chksum; 723 } 724 725 /* 726 * Compute the checksum of a DOS filename for Win95 use 727 */ 728 u_int8_t 729 msdosfs_winChksum(u_int8_t *name) 730 { 731 int i; 732 u_int8_t s; 733 734 for (s = 0, i = 11; --i >= 0; s += *name++) 735 s = (s << 7) | (s >> 1); 736 return s; 737 } 738 739 /* 740 * Determine the number of slots necessary for Win95 names 741 */ 742 int 743 msdosfs_winSlotCnt(const u_char *un, int unlen, int utf8) 744 { 745 const u_char *cp; 746 int len; 747 748 /* 749 * Drop trailing blanks and dots 750 */ 751 for (cp = un + unlen; unlen > 0; unlen--) 752 if (*--cp != ' ' && *cp != '.') 753 break; 754 755 len = utf8 ? utf8ucs2str(un, unlen, NULL, WIN_MAXLEN) : unlen; 756 757 return howmany(len, WIN_CHARS); 758 } 759 760 /* 761 * Scan windows name for characters that must not 762 * appear in a UNIX filename 763 */ 764 static int 765 invalidname(const u_int16_t *in, int n) 766 { 767 while (n-- > 0) { 768 if (*in++ == '/') 769 return 1; 770 } 771 772 return 0; 773 } 774 775 /* 776 * Convert UCS-2 character into UTF-8 777 * return number of output bytes or 0 if output 778 * buffer is too short 779 */ 780 static int 781 ucs2utf8(const u_int16_t *in, u_int8_t *out, int n) 782 { 783 uint16_t inch = le16toh(in[0]); 784 785 if (inch <= 0x007f) { 786 if (n < 1) return 0; 787 if (out) 788 *out++ = inch; 789 return 1; 790 } else if (inch <= 0x07ff) { 791 if (n < 2) return 0; 792 if (out) { 793 *out++ = 0xc0 | (inch >> 6); 794 *out++ = 0x80 | (inch & 0x3f); 795 } 796 return 2; 797 } else { 798 if (n < 3) return 0; 799 if (out) { 800 *out++ = 0xe0 | (inch >> 12); 801 *out++ = 0x80 | ((inch >> 6) & 0x3f); 802 *out++ = 0x80 | (inch & 0x3f); 803 } 804 return 3; 805 } 806 } 807 808 809 /* 810 * Convert UTF-8 bytes into UCS-2 character 811 * return number of input bytes, 0 if input 812 * is too short and -1 if input is invalid 813 */ 814 static int 815 utf8ucs2(const u_int8_t *in, int n, u_int16_t *out) 816 { 817 uint16_t outch; 818 819 if (n < 1) return 0; 820 821 if (in[0] <= 0x7f) { 822 outch = in[0]; 823 if (out) 824 *out = htole16(outch); 825 return 1; 826 } else if (in[0] <= 0xdf) { 827 if (n < 2) return 0; 828 outch = (in[0] & 0x1f) << 6 | (in[1] & 0x3f); 829 if (out) 830 *out = htole16(outch); 831 return 2; 832 } else if (in[0] <= 0xef) { 833 if (n < 3) return 0; 834 outch = (in[0] & 0x1f) << 12 | (in[1] & 0x3f) << 6 | (in[2] & 0x3f); 835 if (out) 836 *out = htole16(outch); 837 return 3; 838 } 839 840 return -1; 841 } 842 843 /* 844 * Convert UCS-2 string into UTF-8 string 845 * return total number of output bytes 846 */ 847 static int 848 ucs2utf8str(const u_int16_t *in, int n, u_int8_t *out, int m) 849 { 850 u_int8_t *p; 851 int outlen; 852 853 p = out; 854 while (n > 0 && *in != 0) { 855 outlen = ucs2utf8(in, out ? p : out, m); 856 if (outlen == 0) 857 break; 858 p += outlen; 859 m -= outlen; 860 in += 1; 861 n -= 1; 862 } 863 864 return p - out; 865 } 866 867 /* 868 * Convert UTF8 string into UCS-2 string 869 * return total number of output characters 870 */ 871 static int 872 utf8ucs2str(const u_int8_t *in, int n, u_int16_t *out, int m) 873 { 874 u_int16_t *p; 875 int inlen; 876 877 p = out; 878 while (n > 0 && *in != 0) { 879 if (m < 1) 880 break; 881 inlen = utf8ucs2(in, n, out ? p : out); 882 if (inlen <= 0) 883 break; 884 in += inlen; 885 n -= inlen; 886 p += 1; 887 m -= 1; 888 } 889 890 return p - out; 891 } 892 893 /* 894 * Convert UCS-2 string into 8bit character string 895 * return total number of output bytes 896 */ 897 static int 898 ucs2char8str(const u_int16_t *in, int n, u_int8_t *out, int m) 899 { 900 u_int8_t *p; 901 u_int16_t inch; 902 903 p = out; 904 while (n > 0 && in[0] != 0) { 905 if (m < 1) 906 break; 907 inch = le16toh(in[0]); 908 if (inch > 255) 909 break; 910 if (p) 911 p[0] = inch; 912 p += 1; 913 m -= 1; 914 in += 1; 915 n -= 1; 916 } 917 918 return p - out; 919 } 920 921 /* 922 * Convert 8bit character string into UCS-2 string 923 * return total number of output characters 924 */ 925 static int 926 char8ucs2str(const u_int8_t *in, int n, u_int16_t *out, int m) 927 { 928 u_int16_t *p; 929 930 p = out; 931 while (n > 0 && in[0] != 0) { 932 if (m < 1) 933 break; 934 if (p) 935 p[0] = htole16(in[0]); 936 p += 1; 937 m -= 1; 938 in += 1; 939 n -= 1; 940 } 941 942 return p - out; 943 } 944 945 static void 946 ucs2pad(u_int16_t *buf, int len, int size) 947 { 948 949 if (len < size-1) 950 buf[len++] = 0x0000; 951 while (len < size) 952 buf[len++] = 0xffff; 953 } 954 955 /* 956 * Fold UCS-2 character to uppercase 957 */ 958 static u_int16_t 959 ucs2fold(u_int16_t w) 960 { 961 int low,high,mid; 962 u_int16_t check; 963 extern const u_int16_t msdosfs_unicode_foldmap[]; 964 extern size_t msdosfs_unicode_foldmap_entries; 965 966 w = le16toh(w); 967 968 low = 0; 969 high = msdosfs_unicode_foldmap_entries / 2; 970 while (low < high) { 971 mid = (low + high)/2; 972 check = msdosfs_unicode_foldmap[2*mid+0]; 973 974 if (w == check) { 975 w = msdosfs_unicode_foldmap[2*mid+1]; 976 break; 977 } 978 979 if (w < check) 980 high = mid; 981 else 982 low = mid+1; 983 } 984 985 w = le16toh(w); 986 987 return w; 988 } 989 990 /* 991 * Compare two UCS-2 strings case-insensitive 992 * 993 * uses the Unicode case folding table 994 */ 995 static int 996 ucs2match(u_int16_t *w1, u_int16_t *w2, int n) 997 { 998 u_int16_t u1, u2; 999 1000 while (n > 0) { 1001 if (*w1 == 0 || *w2 == 0) 1002 return *w1 == *w2; 1003 u1 = ucs2fold(*w1); 1004 u2 = ucs2fold(*w2); 1005 if (u1 != u2) 1006 return 0; 1007 ++w1; 1008 ++w2; 1009 --n; 1010 } 1011 1012 return 1; 1013 } 1014 1015 /* 1016 * Compare two 8bit char conversions case-insensitive 1017 * 1018 * uses the DOS case folding table 1019 */ 1020 static int 1021 char8match(u_int16_t *w1, u_int16_t *w2, int n) 1022 { 1023 u_int16_t u1, u2; 1024 1025 while (n > 0) { 1026 u1 = le16toh(*w1); 1027 u2 = le16toh(*w2); 1028 if (u1 == 0 || u2 == 0) 1029 return u1 == u2; 1030 if (u1 > 255 || u2 > 255) 1031 return 0; 1032 u1 = u2l[u1 & 0xff]; 1033 u2 = u2l[u2 & 0xff]; 1034 if (u1 != u2) 1035 return 0; 1036 ++w1; 1037 ++w2; 1038 --n; 1039 } 1040 1041 return 1; 1042 } 1043 1044