1 /* $FreeBSD$ */ 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/kernel.h> /* defines tz */ 55 #include <sys/systm.h> 56 #include <machine/clock.h> 57 #include <sys/dirent.h> 58 #include <sys/iconv.h> 59 #include <sys/mount.h> 60 61 #include <vfs/msdosfs/bpb.h> 62 #include <vfs/msdosfs/direntry.h> 63 #include <vfs/msdosfs/msdosfsmount.h> 64 65 extern struct iconv_functions *msdos_iconv; 66 67 static int mbsadjpos(const char **, size_t, size_t, int, int, void *); 68 static uint16_t dos2unixchr(u_char *, const u_char **, size_t *, int, 69 struct msdosfsmount *); 70 static uint16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *); 71 static uint16_t win2unixchr(u_char *, uint16_t, struct msdosfsmount *); 72 static uint16_t unix2winchr(const u_char **, size_t *, int, 73 struct msdosfsmount *); 74 75 /* 76 * 0 - character disallowed in long file name. 77 * 1 - character should be replaced by '_' in DOS file name, 78 * and generation number inserted. 79 * 2 - character ('.' and ' ') should be skipped in DOS file name, 80 * and generation number inserted. 81 */ 82 static const u_char 83 unix2dos[256] = { 84 /* iso8859-1 -> cp850 */ 85 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */ 86 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */ 87 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */ 88 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */ 89 2, 0x21, 0, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 90 0x28, 0x29, 0, 1, 1, 0x2d, 2, 0, /* 28-2f */ 91 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 92 0x38, 0x39, 0, 1, 0, 1, 0, 0, /* 38-3f */ 93 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 94 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 95 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 96 0x58, 0x59, 0x5a, 1, 0, 1, 0x5e, 0x5f, /* 58-5f */ 97 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 60-67 */ 98 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 68-6f */ 99 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 70-77 */ 100 0x58, 0x59, 0x5a, 0x7b, 0, 0x7d, 0x7e, 0, /* 78-7f */ 101 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */ 102 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */ 103 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */ 104 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */ 105 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */ 106 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */ 107 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */ 108 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */ 109 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */ 110 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */ 111 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */ 112 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */ 113 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */ 114 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */ 115 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */ 116 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */ 117 }; 118 119 static const u_char 120 dos2unix[256] = { 121 /* cp850 -> iso8859-1 */ 122 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 00-07 */ 123 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 08-0f */ 124 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 10-17 */ 125 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 18-1f */ 126 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 127 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 128 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 129 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 130 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 131 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 132 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 133 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 134 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 135 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 136 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 137 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 138 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */ 139 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */ 140 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */ 141 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, 0x3f, /* 98-9f */ 142 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */ 143 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */ 144 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xc1, 0xc2, 0xc0, /* b0-b7 */ 145 0xa9, 0x3f, 0x3f, 0x3f, 0x3f, 0xa2, 0xa5, 0x3f, /* b8-bf */ 146 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xe3, 0xc3, /* c0-c7 */ 147 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xa4, /* c8-cf */ 148 0xf0, 0xd0, 0xca, 0xcb, 0xc8, 0x3f, 0xcd, 0xce, /* d0-d7 */ 149 0xcf, 0x3f, 0x3f, 0x3f, 0x3f, 0xa6, 0xcc, 0x3f, /* d8-df */ 150 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */ 151 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */ 152 0xad, 0xb1, 0x3f, 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */ 153 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, 0x3f, 0x3f, /* f8-ff */ 154 }; 155 156 static const u_char 157 u2l[256] = { 158 /* tolower */ 159 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 160 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 161 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 162 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 163 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 164 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 165 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 166 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 167 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 168 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 169 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 170 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 171 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 172 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 173 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 174 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 175 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 176 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 177 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 178 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 179 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 180 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 181 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 182 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 183 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 184 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 185 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 186 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 187 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 188 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 189 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 190 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 191 }; 192 193 static const u_char 194 l2u[256] = { 195 /* toupper */ 196 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 197 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 198 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 199 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 200 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 201 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 202 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 203 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 204 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 205 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 206 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 207 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 208 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 209 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 210 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 211 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 212 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 213 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 214 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 215 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 216 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 217 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 218 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 219 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 220 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 221 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 222 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 223 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 224 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 225 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 226 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 227 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 228 }; 229 230 /* 231 * DOS filenames are made of 2 parts, the name part and the extension part. 232 * The name part is 8 characters long and the extension part is 3 233 * characters long. They may contain trailing blanks if the name or 234 * extension are not long enough to fill their respective fields. 235 */ 236 237 /* 238 * Convert a DOS filename to a unix filename. And, return the number of 239 * characters in the resulting unix filename excluding the terminating 240 * null. 241 */ 242 int 243 dos2unixfn(u_char dn[11], u_char *un, int lower, struct msdosfsmount *pmp) 244 { 245 size_t i; 246 int thislong = 0; 247 u_char c, tmpbuf[5]; 248 249 /* 250 * If first char of the filename is SLOT_E5 (0x05), then the real 251 * first char of the filename should be 0xe5. But, they couldn't 252 * just have a 0xe5 mean 0xe5 because that is used to mean a freed 253 * directory slot. Another dos quirk. 254 */ 255 if (*dn == SLOT_E5) 256 *dn = 0xe5; 257 258 /* 259 * Copy the name portion into the unix filename string. 260 */ 261 for (i = 8; i > 0 && *dn != ' ';) { 262 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn), &i, 263 lower & LCASE_BASE, pmp); 264 if (c & 0xff00) { 265 *un++ = c >> 8; 266 thislong++; 267 } 268 *un++ = c; 269 thislong++; 270 } 271 dn += i; 272 273 /* 274 * Now, if there is an extension then put in a period and copy in 275 * the extension. 276 */ 277 if (*dn != ' ') { 278 *un++ = '.'; 279 thislong++; 280 for (i = 3; i > 0 && *dn != ' ';) { 281 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn), 282 &i, lower & LCASE_EXT, pmp); 283 if (c & 0xff00) { 284 *un++ = c >> 8; 285 thislong++; 286 } 287 *un++ = c; 288 thislong++; 289 } 290 } 291 *un++ = 0; 292 293 return (thislong); 294 } 295 296 /* 297 * Convert a unix filename to a DOS filename according to Win95 rules. 298 * If applicable and gen is not 0, it is inserted into the converted 299 * filename as a generation number. 300 * Returns 301 * 0 if name couldn't be converted 302 * 1 if the converted name is the same as the original 303 * (no long filename entry necessary for Win95) 304 * 2 if conversion was successful 305 * 3 if conversion was successful and generation number was inserted 306 */ 307 int 308 unix2dosfn(const u_char *un, u_char dn[12], size_t unlen, u_int gen, 309 struct msdosfsmount *pmp) 310 { 311 ssize_t i, j; 312 int l; 313 int conv = 1; 314 const u_char *cp, *dp, *dp1; 315 u_char gentext[6], *wcp; 316 uint16_t c; 317 318 /* 319 * Fill the dos filename string with blanks. These are DOS's pad 320 * characters. 321 */ 322 for (i = 0; i < 11; i++) 323 dn[i] = ' '; 324 dn[11] = 0; 325 326 /* 327 * The filenames "." and ".." are handled specially, since they 328 * don't follow dos filename rules. 329 */ 330 if (un[0] == '.' && unlen == 1) { 331 dn[0] = '.'; 332 return gen <= 1; 333 } 334 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 335 dn[0] = '.'; 336 dn[1] = '.'; 337 return gen <= 1; 338 } 339 340 /* 341 * Filenames with only blanks and dots are not allowed! 342 */ 343 for (cp = un, i = unlen; --i >= 0; cp++) 344 if (*cp != ' ' && *cp != '.') 345 break; 346 if (i < 0) 347 return 0; 348 349 /* 350 * Filenames with some characters are not allowed! 351 */ 352 for (cp = un, i = unlen; i > 0;) 353 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0) 354 return 0; 355 356 /* 357 * Now find the extension 358 * Note: dot as first char doesn't start extension 359 * and trailing dots and blanks are ignored 360 * Note(2003/7): It seems recent Windows has 361 * defferent rule than this code, that Windows 362 * ignores all dots before extension, and use all 363 * chars as filename except for dots. 364 */ 365 dp = dp1 = NULL; 366 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 367 switch (*cp++) { 368 case '.': 369 if (!dp1) 370 dp1 = cp; 371 break; 372 case ' ': 373 break; 374 default: 375 if (dp1) 376 dp = dp1; 377 dp1 = NULL; 378 break; 379 } 380 } 381 382 /* 383 * Now convert it (this part is for extension). 384 * As Windows XP do, if it's not ascii char, 385 * this function should return 2 or 3, so that checkng out Unicode name. 386 */ 387 if (dp) { 388 if (dp1) 389 l = dp1 - dp; 390 else 391 l = unlen - (dp - un); 392 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) { 393 c = unix2doschr(&cp, (size_t *)&i, pmp); 394 if (c & 0xff00) { 395 dn[j] = c >> 8; 396 if (++j < 11) { 397 dn[j] = c; 398 if (conv != 3) 399 conv = 2; 400 continue; 401 } else { 402 conv = 3; 403 dn[j-1] = ' '; 404 break; 405 } 406 } else { 407 dn[j] = c; 408 } 409 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3) 410 conv = 2; 411 if (dn[j] == 1) { 412 conv = 3; 413 dn[j] = '_'; 414 } 415 if (dn[j] == 2) { 416 conv = 3; 417 dn[j--] = ' '; 418 } 419 } 420 if (i > 0) 421 conv = 3; 422 dp--; 423 } else { 424 for (dp = cp; *--dp == ' ' || *dp == '.';); 425 dp++; 426 } 427 428 /* 429 * Now convert the rest of the name 430 */ 431 for (i = dp - un, j = 0; un < dp && j < 8; j++) { 432 c = unix2doschr(&un, &i, pmp); 433 if (c & 0xff00) { 434 dn[j] = c >> 8; 435 if (++j < 8) { 436 dn[j] = c; 437 if (conv != 3) 438 conv = 2; 439 continue; 440 } else { 441 conv = 3; 442 dn[j-1] = ' '; 443 break; 444 } 445 } else { 446 dn[j] = c; 447 } 448 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3) 449 conv = 2; 450 if (dn[j] == 1) { 451 conv = 3; 452 dn[j] = '_'; 453 } 454 if (dn[j] == 2) { 455 conv = 3; 456 dn[j--] = ' '; 457 } 458 } 459 if (un < dp) 460 conv = 3; 461 /* 462 * If we didn't have any chars in filename, 463 * generate a default 464 */ 465 if (!j) 466 dn[0] = '_'; 467 468 /* 469 * If there wasn't any char dropped, 470 * there is no place for generation numbers 471 */ 472 if (conv != 3) { 473 if (gen > 1) 474 conv = 0; 475 goto done; 476 } 477 478 /* 479 * Now insert the generation number into the filename part 480 */ 481 if (gen == 0) 482 goto done; 483 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 484 *--wcp = gen % 10 + '0'; 485 if (gen) { 486 conv = 0; 487 goto done; 488 } 489 for (i = 8; dn[--i] == ' ';); 490 i++; 491 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 492 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 493 /* 494 * Correct posision to where insert the generation number 495 */ 496 cp = dn; 497 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, 498 pmp->pm_d2u); 499 dn[i++] = '~'; 500 while (wcp < gentext + sizeof(gentext)) 501 dn[i++] = *wcp++; 502 503 /* 504 * Tail of the filename should be space 505 */ 506 while (i < 8) 507 dn[i++] = ' '; 508 conv = 3; 509 510 done: 511 /* 512 * The first character cannot be E5, 513 * because that means a deleted entry 514 */ 515 if (dn[0] == 0xe5) 516 dn[0] = SLOT_E5; 517 518 return conv; 519 } 520 521 /* 522 * Create a Win95 long name directory entry 523 * Note: assumes that the filename is valid, 524 * i.e. doesn't consist solely of blanks and dots 525 */ 526 int 527 unix2winfn(const u_char *un, size_t unlen, struct winentry *wep, int cnt, 528 int chksum, struct msdosfsmount *pmp) 529 { 530 uint8_t *wcp; 531 int i, end; 532 uint16_t code; 533 534 /* 535 * Drop trailing blanks and dots 536 */ 537 unlen = winLenFixup(un, unlen); 538 539 /* 540 * Cut *un for this slot 541 */ 542 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2, 543 pmp->pm_flags, pmp->pm_u2w); 544 545 /* 546 * Initialize winentry to some useful default 547 */ 548 memset(wep, 0xff, sizeof(*wep)); 549 wep->weCnt = cnt; 550 wep->weAttributes = ATTR_WIN95; 551 wep->weReserved1 = 0; 552 wep->weChksum = chksum; 553 wep->weReserved2 = 0; 554 555 /* 556 * Now convert the filename parts 557 */ 558 end = 0; 559 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/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->wePart2, i = sizeof(wep->wePart2)/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 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) { 574 code = unix2winchr(&un, &unlen, 0, pmp); 575 *wcp++ = code; 576 *wcp++ = code >> 8; 577 if (!code) 578 end = WIN_LAST; 579 } 580 if (!unlen) 581 end = WIN_LAST; 582 wep->weCnt |= end; 583 return !end; 584 } 585 586 /* 587 * Compare our filename to the one in the Win95 entry 588 * Returns the checksum or -1 if no match 589 */ 590 int 591 winChkName(struct mbnambuf *nbp, const u_char *un, size_t unlen, int chksum, 592 struct msdosfsmount *pmp) 593 { 594 size_t len; 595 uint16_t c1, c2; 596 u_char *np; 597 struct dirent dirbuf; 598 599 /* 600 * We already have winentry in *nbp. 601 */ 602 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0) 603 return -1; 604 mprintf("winChkName(): un=%s:%zu,d_name=%s:%d\n", 605 un, unlen, dirbuf.d_name, dirbuf.d_namlen); 606 /* 607 * Compare the name parts 608 */ 609 len = dirbuf.d_namlen; 610 if (unlen != len) 611 return -2; 612 613 for (np = dirbuf.d_name; unlen > 0 && len > 0;) { 614 /* 615 * Comparison must be case insensitive, because FAT disallows 616 * to look up or create files in case sensitive even when 617 * it's a long file name. 618 */ 619 c1 = unix2winchr(__DECONST(const u_char **, &np), &len, 620 LCASE_BASE, pmp); 621 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp); 622 if (c1 != c2) 623 return -2; 624 } 625 return chksum; 626 } 627 628 /* 629 * Convert Win95 filename to dirbuf. 630 * Returns the checksum or -1 if impossible 631 */ 632 int 633 win2unixfn(struct mbnambuf *nbp, struct winentry *wep, int chksum, 634 struct msdosfsmount *pmp) 635 { 636 u_char tmpbuf[5]; 637 uint8_t *cp; 638 uint8_t *np, name[WIN_CHARS * 2 + 1]; 639 uint16_t code; 640 int i; 641 642 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 643 || !(wep->weCnt&WIN_CNT)) 644 return -1; 645 646 /* 647 * First compare checksums 648 */ 649 if (wep->weCnt&WIN_LAST) { 650 chksum = wep->weChksum; 651 } else if (chksum != wep->weChksum) 652 chksum = -1; 653 if (chksum == -1) 654 return -1; 655 656 /* 657 * Convert the name parts 658 */ 659 np = name; 660 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) { 661 code = (cp[1] << 8) | cp[0]; 662 switch (code) { 663 case 0: 664 *np = '\0'; 665 if (mbnambuf_write(nbp, name, 666 (wep->weCnt & WIN_CNT) - 1) != 0) 667 return -1; 668 return chksum; 669 case '/': 670 *np = '\0'; 671 return -1; 672 default: 673 code = win2unixchr(tmpbuf, code, pmp); 674 if (code & 0xff00) 675 *np++ = code >> 8; 676 *np++ = code; 677 break; 678 } 679 cp += 2; 680 } 681 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) { 682 code = (cp[1] << 8) | cp[0]; 683 switch (code) { 684 case 0: 685 *np = '\0'; 686 if (mbnambuf_write(nbp, name, 687 (wep->weCnt & WIN_CNT) - 1) != 0) 688 return -1; 689 return chksum; 690 case '/': 691 *np = '\0'; 692 return -1; 693 default: 694 code = win2unixchr(tmpbuf, code, pmp); 695 if (code & 0xff00) 696 *np++ = code >> 8; 697 *np++ = code; 698 break; 699 } 700 cp += 2; 701 } 702 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) { 703 code = (cp[1] << 8) | cp[0]; 704 switch (code) { 705 case 0: 706 *np = '\0'; 707 if (mbnambuf_write(nbp, name, 708 (wep->weCnt & WIN_CNT) - 1) != 0) 709 return -1; 710 return chksum; 711 case '/': 712 *np = '\0'; 713 return -1; 714 default: 715 code = win2unixchr(tmpbuf, code, pmp); 716 if (code & 0xff00) 717 *np++ = code >> 8; 718 *np++ = code; 719 break; 720 } 721 cp += 2; 722 } 723 *np = '\0'; 724 if (mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1) != 0) 725 return -1; 726 return chksum; 727 } 728 729 /* 730 * Compute the unrolled checksum of a DOS filename for Win95 LFN use. 731 */ 732 uint8_t 733 winChksum(uint8_t *name) 734 { 735 int i; 736 uint8_t s; 737 738 for (s = 0, i = 11; --i >= 0; s += *name++) 739 s = (s << 7)|(s >> 1); 740 return (s); 741 } 742 743 /* 744 * Determine the number of slots necessary for Win95 names 745 */ 746 int 747 winSlotCnt(const u_char *un, size_t unlen, struct msdosfsmount *pmp) 748 { 749 size_t wlen; 750 char wn[WIN_MAXLEN * 2 + 1], *wnp; 751 752 unlen = winLenFixup(un, unlen); 753 754 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 755 wlen = WIN_MAXLEN * 2; 756 wnp = wn; 757 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, 758 &wnp, &wlen); 759 if (unlen > 0) 760 return 0; 761 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS); 762 } 763 764 if (unlen > WIN_MAXLEN) 765 return 0; 766 return howmany(unlen, WIN_CHARS); 767 } 768 769 /* 770 * Determine the number of bytes necessary for Win95 names 771 */ 772 size_t 773 winLenFixup(const u_char *un, size_t unlen) 774 { 775 for (un += unlen; unlen > 0; unlen--) 776 if (*--un != ' ' && *un != '.') 777 break; 778 return unlen; 779 } 780 781 /* 782 * Store an area with multi byte string instr, and returns left 783 * byte of instr and moves pointer forward. The area's size is 784 * inlen or outlen. 785 */ 786 static int 787 mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, 788 void *handle) 789 { 790 char *outp, outstr[outlen * weight + 1]; 791 792 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) { 793 outp = outstr; 794 outlen *= weight; 795 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen); 796 return (inlen); 797 } 798 799 (*instr) += min(inlen, outlen); 800 return (inlen - min(inlen, outlen)); 801 } 802 803 /* 804 * Convert DOS char to Local char 805 */ 806 static uint16_t 807 dos2unixchr(u_char *outbuf, const u_char **instr, size_t *ilen, int lower, 808 struct msdosfsmount *pmp) 809 { 810 u_char c; 811 char *outp; 812 uint16_t wc; 813 size_t len, olen; 814 815 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 816 olen = len = 2; 817 outp = outbuf; 818 819 if (lower & (LCASE_BASE | LCASE_EXT)) 820 msdosfs_iconv->convchr_case(pmp->pm_d2u, 821 (const char **)instr, ilen, 822 &outp, &olen, KICONV_LOWER); 823 else 824 msdosfs_iconv->convchr(pmp->pm_d2u, 825 (const char **)instr, 826 ilen, &outp, &olen); 827 len -= olen; 828 829 /* 830 * return '?' if failed to convert 831 */ 832 if (len == 0) { 833 (*ilen)--; 834 (*instr)++; 835 return ('?'); 836 } 837 wc = 0; 838 while(len--) 839 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 840 return (wc); 841 } 842 (*ilen)--; 843 c = *(*instr)++; 844 c = dos2unix[c]; 845 if (lower & (LCASE_BASE | LCASE_EXT)) 846 c = u2l[c]; 847 return ((uint16_t)c); 848 } 849 850 /* 851 * Convert Local char to DOS char 852 */ 853 static uint16_t 854 unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp) 855 { 856 u_char c; 857 char *up, *outp, unicode[3], outbuf[3]; 858 uint16_t wc; 859 size_t len, ucslen, unixlen, olen; 860 861 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 862 /* 863 * to hide an invisible character, using a unicode filter 864 */ 865 ucslen = 2; 866 len = *ilen; 867 up = unicode; 868 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 869 ilen, &up, &ucslen); 870 unixlen = len - *ilen; 871 872 /* 873 * cannot be converted 874 */ 875 if (unixlen == 0) { 876 (*ilen)--; 877 (*instr)++; 878 return (0); 879 } 880 881 /* 882 * return magic number for ascii char 883 */ 884 if (unixlen == 1) { 885 c = *(*instr -1); 886 if (! (c & 0x80)) { 887 c = unix2dos[c]; 888 if (c <= 2) 889 return (c); 890 } 891 } 892 893 /* 894 * now convert using libiconv 895 */ 896 *instr -= unixlen; 897 *ilen = len; 898 899 olen = len = 2; 900 outp = outbuf; 901 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr, 902 ilen, &outp, &olen, 903 KICONV_FROM_UPPER); 904 len -= olen; 905 906 /* 907 * cannot be converted, but has unicode char, should return magic number 908 */ 909 if (len == 0) { 910 (*ilen) -= unixlen; 911 (*instr) += unixlen; 912 return (1); 913 } 914 915 wc = 0; 916 while(len--) 917 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 918 return (wc); 919 } 920 921 (*ilen)--; 922 c = *(*instr)++; 923 c = l2u[c]; 924 c = unix2dos[c]; 925 return ((uint16_t)c); 926 } 927 928 /* 929 * Convert Windows char to Local char 930 */ 931 static uint16_t 932 win2unixchr(u_char *outbuf, uint16_t wc, struct msdosfsmount *pmp) 933 { 934 u_char *inp, *outp, inbuf[3]; 935 size_t ilen, olen, len; 936 937 if (wc == 0) 938 return (0); 939 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 940 inbuf[0] = (u_char)(wc>>8); 941 inbuf[1] = (u_char)wc; 942 inbuf[2] = '\0'; 943 ilen = olen = len = 2; 944 inp = inbuf; 945 outp = outbuf; 946 msdosfs_iconv->convchr(pmp->pm_w2u, __DECONST(const char **, 947 &inp), &ilen, (char **)&outp, &olen); 948 len -= olen; 949 950 /* 951 * return '?' if failed to convert 952 */ 953 if (len == 0) { 954 wc = '?'; 955 return (wc); 956 } 957 wc = 0; 958 while(len--) 959 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 960 return (wc); 961 } 962 if (wc & 0xff00) 963 wc = '?'; 964 return (wc); 965 } 966 967 /* 968 * Convert Local char to Windows char 969 */ 970 static uint16_t 971 unix2winchr(const u_char **instr, size_t *ilen, int lower, 972 struct msdosfsmount *pmp) 973 { 974 u_char *outp, outbuf[3]; 975 uint16_t wc; 976 size_t olen; 977 978 if (*ilen == 0) 979 return (0); 980 981 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 982 outp = outbuf; 983 olen = 2; 984 if (lower & (LCASE_BASE | LCASE_EXT)) 985 msdosfs_iconv->convchr_case(pmp->pm_u2w, 986 (const char **)instr, 987 ilen, (char **)&outp, &olen, 988 KICONV_FROM_LOWER); 989 else 990 msdosfs_iconv->convchr(pmp->pm_u2w, 991 (const char **)instr, 992 ilen, (char **)&outp, &olen); 993 /* 994 * return '0' if end of filename 995 */ 996 if (olen == 2) 997 return (0); 998 999 wc = (outbuf[0]<<8) | outbuf[1]; 1000 return (wc); 1001 } 1002 1003 (*ilen)--; 1004 wc = (*instr)[0]; 1005 if (lower & (LCASE_BASE | LCASE_EXT)) 1006 wc = u2l[wc]; 1007 (*instr)++; 1008 return (wc); 1009 } 1010 1011 /* 1012 * Initialize the temporary concatenation buffer. 1013 */ 1014 void 1015 mbnambuf_init(struct mbnambuf *nbp) 1016 { 1017 nbp->nb_len = 0; 1018 nbp->nb_last_id = -1; 1019 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0'; 1020 } 1021 1022 /* 1023 * Fill out our concatenation buffer with the given substring, at the offset 1024 * specified by its id. Since this function must be called with ids in 1025 * descending order, we take advantage of the fact that ASCII substrings are 1026 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all 1027 * previous (i.e. higher id) substrings upwards to make room for this one. 1028 * This only penalizes portions of substrings that contain more than 1029 * WIN_CHARS bytes when they are first encountered. 1030 */ 1031 int 1032 mbnambuf_write(struct mbnambuf *nbp, char *name, int id) 1033 { 1034 char *slot; 1035 size_t count, newlen; 1036 1037 if (nbp->nb_len != 0 && id != nbp->nb_last_id - 1) { 1038 kprintf("msdosfs: non-decreasing id: id %d, last id %d\n", 1039 id, nbp->nb_last_id); 1040 return (EINVAL); 1041 } 1042 1043 /* Will store this substring in a WIN_CHARS-aligned slot. */ 1044 slot = &nbp->nb_buf[id * WIN_CHARS]; 1045 count = strlen(name); 1046 newlen = nbp->nb_len + count; 1047 if (newlen > WIN_MAXLEN || newlen > 127) { 1048 kprintf("msdosfs: file name length %zu too large\n", newlen); 1049 return (ENAMETOOLONG); 1050 } 1051 1052 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */ 1053 if (count > WIN_CHARS && nbp->nb_len != 0) { 1054 if ((id * WIN_CHARS + count + nbp->nb_len) > 1055 sizeof(nbp->nb_buf)) 1056 return (ENAMETOOLONG); 1057 1058 memmove(slot + count, slot + WIN_CHARS, nbp->nb_len); 1059 } 1060 1061 /* Copy in the substring to its slot and update length so far. */ 1062 memcpy(slot, name, count); 1063 nbp->nb_len = newlen; 1064 nbp->nb_last_id = id; 1065 1066 return (0); 1067 } 1068 1069 /* 1070 * Take the completed string and use it to setup the struct dirent. 1071 * Be sure to always nul-terminate the d_name and then copy the string 1072 * from our buffer. Note that this function assumes the full string has 1073 * been reassembled in the buffer. If it's called before all substrings 1074 * have been written via mbnambuf_write(), the result will be incorrect. 1075 */ 1076 char * 1077 mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp) 1078 { 1079 if (nbp->nb_len > sizeof(dp->d_name) - 1) { 1080 mbnambuf_init(nbp); 1081 return (NULL); 1082 } 1083 memcpy(dp->d_name, &nbp->nb_buf[0], nbp->nb_len); 1084 dp->d_name[nbp->nb_len] = '\0'; 1085 dp->d_namlen = nbp->nb_len; 1086 1087 mbnambuf_init(nbp); 1088 return (dp->d_name); 1089 } 1090