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