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/systm.h> 55 #include <sys/dirent.h> 56 #include <sys/iconv.h> 57 #include <sys/mount.h> 58 59 #include <fs/msdosfs/bpb.h> 60 #include <fs/msdosfs/direntry.h> 61 #include <fs/msdosfs/msdosfsmount.h> 62 63 extern struct iconv_functions *msdosfs_iconv; 64 65 static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle); 66 static u_char * dos2unixchr(u_char *, const u_char **, size_t *, int, struct msdosfsmount *); 67 static uint16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *); 68 static u_char * win2unixchr(u_char *, uint16_t, struct msdosfsmount *); 69 static uint16_t unix2winchr(const u_char **, size_t *, int, struct msdosfsmount *); 70 71 /* 72 * 0 - character disallowed in long file name. 73 * 1 - character should be replaced by '_' in DOS file name, 74 * and generation number inserted. 75 * 2 - character ('.' and ' ') should be skipped in DOS file name, 76 * and generation number inserted. 77 */ 78 static const u_char 79 unix2dos[256] = { 80 /* iso8859-1 -> cp850 */ 81 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */ 82 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */ 83 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */ 84 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */ 85 2, 0x21, 0, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 86 0x28, 0x29, 0, 1, 1, 0x2d, 2, 0, /* 28-2f */ 87 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 88 0x38, 0x39, 0, 1, 0, 1, 0, 0, /* 38-3f */ 89 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 90 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 91 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 92 0x58, 0x59, 0x5a, 1, 0, 1, 0x5e, 0x5f, /* 58-5f */ 93 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 60-67 */ 94 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 68-6f */ 95 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 70-77 */ 96 0x58, 0x59, 0x5a, 0x7b, 0, 0x7d, 0x7e, 0, /* 78-7f */ 97 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */ 98 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */ 99 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */ 100 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */ 101 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */ 102 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */ 103 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */ 104 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */ 105 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */ 106 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */ 107 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */ 108 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */ 109 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */ 110 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */ 111 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */ 112 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */ 113 }; 114 115 static const u_char 116 dos2unix[256] = { 117 /* cp850 -> iso8859-1 */ 118 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 00-07 */ 119 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 08-0f */ 120 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 10-17 */ 121 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 18-1f */ 122 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 123 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 124 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 125 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 126 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 127 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 128 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 129 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 130 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 131 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 132 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 133 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 134 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */ 135 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */ 136 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */ 137 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, 0x3f, /* 98-9f */ 138 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */ 139 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */ 140 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xc1, 0xc2, 0xc0, /* b0-b7 */ 141 0xa9, 0x3f, 0x3f, 0x3f, 0x3f, 0xa2, 0xa5, 0x3f, /* b8-bf */ 142 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xe3, 0xc3, /* c0-c7 */ 143 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xa4, /* c8-cf */ 144 0xf0, 0xd0, 0xca, 0xcb, 0xc8, 0x3f, 0xcd, 0xce, /* d0-d7 */ 145 0xcf, 0x3f, 0x3f, 0x3f, 0x3f, 0xa6, 0xcc, 0x3f, /* d8-df */ 146 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */ 147 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */ 148 0xad, 0xb1, 0x3f, 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */ 149 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, 0x3f, 0x3f, /* f8-ff */ 150 }; 151 152 static const u_char 153 u2l[256] = { 154 /* tolower */ 155 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 156 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 157 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 158 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 159 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 160 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 161 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 162 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 163 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 164 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 165 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 166 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 167 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 168 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 169 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 170 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 171 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 172 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 173 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 174 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 175 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 176 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 177 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 178 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 179 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 180 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 181 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 182 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 183 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 184 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 185 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 186 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 187 }; 188 189 static const u_char 190 l2u[256] = { 191 /* toupper */ 192 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 193 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 194 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 195 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 196 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 197 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 198 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 199 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 200 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 201 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 202 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 203 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 204 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 205 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 206 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 207 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 208 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 209 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 210 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 211 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 212 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 213 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 214 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 215 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 216 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 217 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 218 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 219 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 220 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 221 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 222 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 223 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 224 }; 225 226 /* 227 * DOS filenames are made of 2 parts, the name part and the extension part. 228 * The name part is 8 characters long and the extension part is 3 229 * characters long. They may contain trailing blanks if the name or 230 * extension are not long enough to fill their respective fields. 231 */ 232 233 /* 234 * Convert a DOS filename to a unix filename. And, return the number of 235 * characters in the resulting unix filename excluding the terminating 236 * null. 237 */ 238 int 239 dos2unixfn(u_char dn[11], u_char *un, int lower, struct msdosfsmount *pmp) 240 { 241 size_t i; 242 int thislong = 0; 243 u_char *c, tmpbuf[5]; 244 245 /* 246 * If first char of the filename is SLOT_E5 (0x05), then the real 247 * first char of the filename should be 0xe5. But, they couldn't 248 * just have a 0xe5 mean 0xe5 because that is used to mean a freed 249 * directory slot. Another dos quirk. 250 */ 251 if (*dn == SLOT_E5) 252 *dn = 0xe5; 253 254 /* 255 * Copy the name portion into the unix filename string. 256 */ 257 for (i = 8; i > 0 && *dn != ' ';) { 258 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn), &i, 259 lower & LCASE_BASE, pmp); 260 while (*c != '\0') { 261 *un++ = *c++; 262 thislong++; 263 } 264 } 265 dn += i; 266 267 /* 268 * Now, if there is an extension then put in a period and copy in 269 * the extension. 270 */ 271 if (*dn != ' ') { 272 *un++ = '.'; 273 thislong++; 274 for (i = 3; i > 0 && *dn != ' ';) { 275 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn), 276 &i, lower & LCASE_EXT, pmp); 277 while (*c != '\0') { 278 *un++ = *c++; 279 thislong++; 280 } 281 } 282 } 283 *un++ = 0; 284 285 return (thislong); 286 } 287 288 /* 289 * Convert a unix filename to a DOS filename according to Win95 rules. 290 * If applicable and gen is not 0, it is inserted into the converted 291 * filename as a generation number. 292 * Returns 293 * 0 if name couldn't be converted 294 * 1 if the converted name is the same as the original 295 * (no long filename entry necessary for Win95) 296 * 2 if conversion was successful 297 * 3 if conversion was successful and generation number was inserted 298 */ 299 int 300 unix2dosfn(const u_char *un, u_char dn[12], size_t unlen, u_int gen, 301 struct msdosfsmount *pmp) 302 { 303 ssize_t i, j; 304 int l; 305 int conv = 1; 306 const u_char *cp, *dp, *dp1; 307 u_char gentext[6], *wcp; 308 uint16_t c; 309 310 /* 311 * Fill the dos filename string with blanks. These are DOS's pad 312 * characters. 313 */ 314 for (i = 0; i < 11; i++) 315 dn[i] = ' '; 316 dn[11] = 0; 317 318 /* 319 * The filenames "." and ".." are handled specially, since they 320 * don't follow dos filename rules. 321 */ 322 if (un[0] == '.' && unlen == 1) { 323 dn[0] = '.'; 324 return gen <= 1; 325 } 326 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 327 dn[0] = '.'; 328 dn[1] = '.'; 329 return gen <= 1; 330 } 331 332 /* 333 * Filenames with only blanks and dots are not allowed! 334 */ 335 for (cp = un, i = unlen; --i >= 0; cp++) 336 if (*cp != ' ' && *cp != '.') 337 break; 338 if (i < 0) 339 return 0; 340 341 /* 342 * Filenames with some characters are not allowed! 343 */ 344 for (cp = un, i = unlen; i > 0;) 345 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0) 346 return 0; 347 348 /* 349 * Now find the extension 350 * Note: dot as first char doesn't start extension 351 * and trailing dots and blanks are ignored 352 * Note(2003/7): It seems recent Windows has 353 * defferent rule than this code, that Windows 354 * ignores all dots before extension, and use all 355 * chars as filename except for dots. 356 */ 357 dp = dp1 = NULL; 358 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 359 switch (*cp++) { 360 case '.': 361 if (!dp1) 362 dp1 = cp; 363 break; 364 case ' ': 365 break; 366 default: 367 if (dp1) 368 dp = dp1; 369 dp1 = NULL; 370 break; 371 } 372 } 373 374 /* 375 * Now convert it (this part is for extension). 376 * As Windows XP do, if it's not ascii char, 377 * this function should return 2 or 3, so that checkng out Unicode name. 378 */ 379 if (dp) { 380 if (dp1) 381 l = dp1 - dp; 382 else 383 l = unlen - (dp - un); 384 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) { 385 c = unix2doschr(&cp, (size_t *)&i, pmp); 386 if (c & 0xff00) { 387 dn[j] = c >> 8; 388 if (++j < 11) { 389 dn[j] = c; 390 if (conv != 3) 391 conv = 2; 392 continue; 393 } else { 394 conv = 3; 395 dn[j-1] = ' '; 396 break; 397 } 398 } else { 399 dn[j] = c; 400 } 401 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3) 402 conv = 2; 403 if (dn[j] == 1) { 404 conv = 3; 405 dn[j] = '_'; 406 } 407 if (dn[j] == 2) { 408 conv = 3; 409 dn[j--] = ' '; 410 } 411 } 412 if (i > 0) 413 conv = 3; 414 dp--; 415 } else { 416 for (dp = cp; *--dp == ' ' || *dp == '.';); 417 dp++; 418 } 419 420 /* 421 * Now convert the rest of the name 422 */ 423 for (i = dp - un, j = 0; un < dp && j < 8; j++) { 424 c = unix2doschr(&un, &i, pmp); 425 if (c & 0xff00) { 426 dn[j] = c >> 8; 427 if (++j < 8) { 428 dn[j] = c; 429 if (conv != 3) 430 conv = 2; 431 continue; 432 } else { 433 conv = 3; 434 dn[j-1] = ' '; 435 break; 436 } 437 } else { 438 dn[j] = c; 439 } 440 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3) 441 conv = 2; 442 if (dn[j] == 1) { 443 conv = 3; 444 dn[j] = '_'; 445 } 446 if (dn[j] == 2) { 447 conv = 3; 448 dn[j--] = ' '; 449 } 450 } 451 if (un < dp) 452 conv = 3; 453 /* 454 * If we didn't have any chars in filename, 455 * generate a default 456 */ 457 if (!j) 458 dn[0] = '_'; 459 460 /* 461 * If there wasn't any char dropped, 462 * there is no place for generation numbers 463 */ 464 if (conv != 3) { 465 if (gen > 1) 466 conv = 0; 467 goto done; 468 } 469 470 /* 471 * Now insert the generation number into the filename part 472 */ 473 if (gen == 0) 474 goto done; 475 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 476 *--wcp = gen % 10 + '0'; 477 if (gen) { 478 conv = 0; 479 goto done; 480 } 481 for (i = 8; dn[--i] == ' ';); 482 i++; 483 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 484 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 485 /* 486 * Correct posision to where insert the generation number 487 */ 488 cp = dn; 489 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u); 490 491 dn[i++] = '~'; 492 while (wcp < gentext + sizeof(gentext)) 493 dn[i++] = *wcp++; 494 495 /* 496 * Tail of the filename should be space 497 */ 498 while (i < 8) 499 dn[i++] = ' '; 500 conv = 3; 501 502 done: 503 /* 504 * The first character cannot be E5, 505 * because that means a deleted entry 506 */ 507 if (dn[0] == 0xe5) 508 dn[0] = SLOT_E5; 509 510 return conv; 511 } 512 513 /* 514 * Create a Win95 long name directory entry 515 * Note: assumes that the filename is valid, 516 * i.e. doesn't consist solely of blanks and dots 517 */ 518 int 519 unix2winfn(const u_char *un, size_t unlen, struct winentry *wep, int cnt, 520 int chksum, struct msdosfsmount *pmp) 521 { 522 uint8_t *wcp; 523 int i, end; 524 uint16_t code; 525 526 /* 527 * Drop trailing blanks and dots 528 */ 529 unlen = winLenFixup(un, unlen); 530 531 /* 532 * Cut *un for this slot 533 */ 534 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2, 535 pmp->pm_flags, pmp->pm_u2w); 536 537 /* 538 * Initialize winentry to some useful default 539 */ 540 memset(wep, 0xff, sizeof(*wep)); 541 wep->weCnt = cnt; 542 wep->weAttributes = ATTR_WIN95; 543 wep->weReserved1 = 0; 544 wep->weChksum = chksum; 545 wep->weReserved2 = 0; 546 547 /* 548 * Now convert the filename parts 549 */ 550 end = 0; 551 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) { 552 code = unix2winchr(&un, &unlen, 0, pmp); 553 *wcp++ = code; 554 *wcp++ = code >> 8; 555 if (!code) 556 end = WIN_LAST; 557 } 558 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) { 559 code = unix2winchr(&un, &unlen, 0, pmp); 560 *wcp++ = code; 561 *wcp++ = code >> 8; 562 if (!code) 563 end = WIN_LAST; 564 } 565 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) { 566 code = unix2winchr(&un, &unlen, 0, pmp); 567 *wcp++ = code; 568 *wcp++ = code >> 8; 569 if (!code) 570 end = WIN_LAST; 571 } 572 if (!unlen) 573 end = WIN_LAST; 574 wep->weCnt |= end; 575 return !end; 576 } 577 578 /* 579 * Compare our filename to the one in the Win95 entry 580 * Returns the checksum or -1 if no match 581 */ 582 int 583 winChkName(struct mbnambuf *nbp, const u_char *un, size_t unlen, int chksum, 584 struct msdosfsmount *pmp) 585 { 586 size_t len; 587 uint16_t c1, c2; 588 u_char *np; 589 struct dirent dirbuf; 590 591 /* 592 * We already have winentry in *nbp. 593 */ 594 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0) 595 return -1; 596 597 #ifdef MSDOSFS_DEBUG 598 printf("winChkName(): un=%s:%zu,d_name=%s:%d\n", un, unlen, 599 dirbuf.d_name, 600 dirbuf.d_namlen); 601 #endif 602 603 /* 604 * Compare the name parts 605 */ 606 len = dirbuf.d_namlen; 607 if (unlen != len) 608 return -2; 609 610 for (np = dirbuf.d_name; unlen > 0 && len > 0;) { 611 /* 612 * Comparison must be case insensitive, because FAT disallows 613 * to look up or create files in case sensitive even when 614 * it's a long file name. 615 */ 616 c1 = unix2winchr(__DECONST(const u_char **, &np), &len, 617 LCASE_BASE, pmp); 618 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp); 619 if (c1 != c2) 620 return -2; 621 } 622 return chksum; 623 } 624 625 /* 626 * Convert Win95 filename to dirbuf. 627 * Returns the checksum or -1 if impossible 628 */ 629 int 630 win2unixfn(struct mbnambuf *nbp, struct winentry *wep, int chksum, 631 struct msdosfsmount *pmp) 632 { 633 u_char *c, tmpbuf[5]; 634 uint8_t *cp; 635 uint8_t *np, name[WIN_CHARS * 3 + 1]; 636 uint16_t code; 637 int i; 638 639 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 640 || !(wep->weCnt&WIN_CNT)) 641 return -1; 642 643 /* 644 * First compare checksums 645 */ 646 if (wep->weCnt&WIN_LAST) { 647 chksum = wep->weChksum; 648 } else if (chksum != wep->weChksum) 649 chksum = -1; 650 if (chksum == -1) 651 return -1; 652 653 /* 654 * Convert the name parts 655 */ 656 np = name; 657 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) { 658 code = (cp[1] << 8) | cp[0]; 659 switch (code) { 660 case 0: 661 *np = '\0'; 662 if (mbnambuf_write(nbp, name, 663 (wep->weCnt & WIN_CNT) - 1) != 0) 664 return -1; 665 return chksum; 666 case '/': 667 *np = '\0'; 668 return -1; 669 default: 670 c = win2unixchr(tmpbuf, code, pmp); 671 while (*c != '\0') 672 *np++ = *c++; 673 break; 674 } 675 cp += 2; 676 } 677 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) { 678 code = (cp[1] << 8) | cp[0]; 679 switch (code) { 680 case 0: 681 *np = '\0'; 682 if (mbnambuf_write(nbp, name, 683 (wep->weCnt & WIN_CNT) - 1) != 0) 684 return -1; 685 return chksum; 686 case '/': 687 *np = '\0'; 688 return -1; 689 default: 690 c = win2unixchr(tmpbuf, code, pmp); 691 while (*c != '\0') 692 *np++ = *c++; 693 break; 694 } 695 cp += 2; 696 } 697 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) { 698 code = (cp[1] << 8) | cp[0]; 699 switch (code) { 700 case 0: 701 *np = '\0'; 702 if (mbnambuf_write(nbp, name, 703 (wep->weCnt & WIN_CNT) - 1) != 0) 704 return -1; 705 return chksum; 706 case '/': 707 *np = '\0'; 708 return -1; 709 default: 710 c = win2unixchr(tmpbuf, code, pmp); 711 while (*c != '\0') 712 *np++ = *c++; 713 break; 714 } 715 cp += 2; 716 } 717 *np = '\0'; 718 if (mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1) != 0) 719 return -1; 720 return chksum; 721 } 722 723 /* 724 * Compute the unrolled checksum of a DOS filename for Win95 LFN use. 725 */ 726 uint8_t 727 winChksum(uint8_t *name) 728 { 729 int i; 730 uint8_t s; 731 732 for (s = 0, i = 11; --i >= 0; s += *name++) 733 s = (s << 7)|(s >> 1); 734 return (s); 735 } 736 737 /* 738 * Determine the number of slots necessary for Win95 names 739 */ 740 int 741 winSlotCnt(const u_char *un, size_t unlen, struct msdosfsmount *pmp) 742 { 743 size_t wlen; 744 char wn[WIN_MAXLEN * 2 + 1], *wnp; 745 746 unlen = winLenFixup(un, unlen); 747 748 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 749 wlen = WIN_MAXLEN * 2; 750 wnp = wn; 751 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen); 752 if (unlen > 0) 753 return 0; 754 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS); 755 } 756 757 if (unlen > WIN_MAXLEN) 758 return 0; 759 return howmany(unlen, WIN_CHARS); 760 } 761 762 /* 763 * Determine the number of bytes necessary for Win95 names 764 */ 765 size_t 766 winLenFixup(const u_char *un, size_t unlen) 767 { 768 for (un += unlen; unlen > 0; unlen--) 769 if (*--un != ' ' && *un != '.') 770 break; 771 return unlen; 772 } 773 774 /* 775 * Store an area with multi byte string instr, and returns left 776 * byte of instr and moves pointer forward. The area's size is 777 * inlen or outlen. 778 */ 779 static int 780 mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle) 781 { 782 char *outp, outstr[outlen * weight + 1]; 783 784 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) { 785 outp = outstr; 786 outlen *= weight; 787 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen); 788 return (inlen); 789 } 790 791 (*instr) += min(inlen, outlen); 792 return (inlen - min(inlen, outlen)); 793 } 794 795 /* 796 * Convert DOS char to Local char 797 */ 798 static u_char * 799 dos2unixchr(u_char *outbuf, const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 800 { 801 u_char c, *outp; 802 size_t len, olen; 803 804 outp = outbuf; 805 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 806 olen = len = 4; 807 808 if (lower & (LCASE_BASE | LCASE_EXT)) 809 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr, 810 ilen, (char **)&outp, &olen, KICONV_LOWER); 811 else 812 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr, 813 ilen, (char **)&outp, &olen); 814 len -= olen; 815 816 /* 817 * return '?' if failed to convert 818 */ 819 if (len == 0) { 820 (*ilen)--; 821 (*instr)++; 822 *outp++ = '?'; 823 } 824 } else { 825 (*ilen)--; 826 c = *(*instr)++; 827 c = dos2unix[c]; 828 if (lower & (LCASE_BASE | LCASE_EXT)) 829 c = u2l[c]; 830 *outp++ = c; 831 outbuf[1] = '\0'; 832 } 833 834 *outp = '\0'; 835 outp = outbuf; 836 return (outp); 837 } 838 839 /* 840 * Convert Local char to DOS char 841 */ 842 static uint16_t 843 unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp) 844 { 845 u_char c; 846 char *up, *outp, unicode[3], outbuf[3]; 847 uint16_t wc; 848 size_t len, ucslen, unixlen, olen; 849 850 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 851 /* 852 * to hide an invisible character, using a unicode filter 853 */ 854 ucslen = 2; 855 len = *ilen; 856 up = unicode; 857 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 858 ilen, &up, &ucslen); 859 unixlen = len - *ilen; 860 861 /* 862 * cannot be converted 863 */ 864 if (unixlen == 0) { 865 (*ilen)--; 866 (*instr)++; 867 return (0); 868 } 869 870 /* 871 * return magic number for ascii char 872 */ 873 if (unixlen == 1) { 874 c = *(*instr -1); 875 if (! (c & 0x80)) { 876 c = unix2dos[c]; 877 if (c <= 2) 878 return (c); 879 } 880 } 881 882 /* 883 * now convert using libiconv 884 */ 885 *instr -= unixlen; 886 *ilen = len; 887 888 olen = len = 2; 889 outp = outbuf; 890 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr, 891 ilen, &outp, &olen, KICONV_FROM_UPPER); 892 len -= olen; 893 894 /* 895 * cannot be converted, but has unicode char, should return magic number 896 */ 897 if (len == 0) { 898 (*ilen) -= unixlen; 899 (*instr) += unixlen; 900 return (1); 901 } 902 903 wc = 0; 904 while(len--) 905 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 906 return (wc); 907 } 908 909 (*ilen)--; 910 c = *(*instr)++; 911 c = l2u[c]; 912 c = unix2dos[c]; 913 return ((uint16_t)c); 914 } 915 916 /* 917 * Convert Windows char to Local char 918 */ 919 static u_char * 920 win2unixchr(u_char *outbuf, uint16_t wc, struct msdosfsmount *pmp) 921 { 922 u_char *inp, *outp, inbuf[3]; 923 size_t ilen, olen, len; 924 925 outp = outbuf; 926 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 927 inbuf[0] = (u_char)(wc>>8); 928 inbuf[1] = (u_char)wc; 929 inbuf[2] = '\0'; 930 931 ilen = 2; 932 olen = len = 4; 933 inp = inbuf; 934 msdosfs_iconv->convchr(pmp->pm_w2u, __DECONST(const char **, 935 &inp), &ilen, (char **)&outp, &olen); 936 len -= olen; 937 938 /* 939 * return '?' if failed to convert 940 */ 941 if (len == 0) 942 *outp++ = '?'; 943 } else { 944 *outp++ = (wc & 0xff00) ? '?' : (u_char)(wc & 0xff); 945 } 946 947 *outp = '\0'; 948 outp = outbuf; 949 return (outp); 950 } 951 952 /* 953 * Convert Local char to Windows char 954 */ 955 static uint16_t 956 unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 957 { 958 u_char *outp, outbuf[3]; 959 uint16_t wc; 960 size_t olen; 961 962 if (*ilen == 0) 963 return (0); 964 965 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 966 outp = outbuf; 967 olen = 2; 968 if (lower & (LCASE_BASE | LCASE_EXT)) 969 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr, 970 ilen, (char **)&outp, &olen, 971 KICONV_FROM_LOWER); 972 else 973 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 974 ilen, (char **)&outp, &olen); 975 976 /* 977 * return '0' if end of filename 978 */ 979 if (olen == 2) 980 return (0); 981 982 wc = (outbuf[0]<<8) | outbuf[1]; 983 984 return (wc); 985 } 986 987 (*ilen)--; 988 wc = (*instr)[0]; 989 if (lower & (LCASE_BASE | LCASE_EXT)) 990 wc = u2l[wc]; 991 (*instr)++; 992 return (wc); 993 } 994 995 /* 996 * Initialize the temporary concatenation buffer. 997 */ 998 void 999 mbnambuf_init(struct mbnambuf *nbp) 1000 { 1001 1002 nbp->nb_len = 0; 1003 nbp->nb_last_id = -1; 1004 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0'; 1005 } 1006 1007 /* 1008 * Fill out our concatenation buffer with the given substring, at the offset 1009 * specified by its id. Since this function must be called with ids in 1010 * descending order, we take advantage of the fact that ASCII substrings are 1011 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all 1012 * previous (i.e. higher id) substrings upwards to make room for this one. 1013 * This only penalizes portions of substrings that contain more than 1014 * WIN_CHARS bytes when they are first encountered. 1015 */ 1016 int 1017 mbnambuf_write(struct mbnambuf *nbp, char *name, int id) 1018 { 1019 char *slot; 1020 size_t count, newlen; 1021 1022 if (nbp->nb_len != 0 && id != nbp->nb_last_id - 1) { 1023 #ifdef MSDOSFS_DEBUG 1024 printf("msdosfs: non-decreasing id: id %d, last id %d\n", 1025 id, nbp->nb_last_id); 1026 #endif 1027 return (EINVAL); 1028 } 1029 1030 /* Will store this substring in a WIN_CHARS-aligned slot. */ 1031 slot = &nbp->nb_buf[id * WIN_CHARS]; 1032 count = strlen(name); 1033 newlen = nbp->nb_len + count; 1034 if (newlen > WIN_MAXLEN || newlen > MAXNAMLEN) { 1035 #ifdef MSDOSFS_DEBUG 1036 printf("msdosfs: file name length %zu too large\n", newlen); 1037 #endif 1038 return (ENAMETOOLONG); 1039 } 1040 1041 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */ 1042 if (count > WIN_CHARS && nbp->nb_len != 0) { 1043 if ((id * WIN_CHARS + count + nbp->nb_len) > 1044 sizeof(nbp->nb_buf)) 1045 return (ENAMETOOLONG); 1046 1047 memmove(slot + count, slot + WIN_CHARS, nbp->nb_len); 1048 } 1049 1050 /* Copy in the substring to its slot and update length so far. */ 1051 memcpy(slot, name, count); 1052 nbp->nb_len = newlen; 1053 nbp->nb_last_id = id; 1054 1055 return (0); 1056 } 1057 1058 /* 1059 * Take the completed string and use it to setup the struct dirent. 1060 * Be sure to always nul-terminate the d_name and then copy the string 1061 * from our buffer. Note that this function assumes the full string has 1062 * been reassembled in the buffer. If it's called before all substrings 1063 * have been written via mbnambuf_write(), the result will be incorrect. 1064 */ 1065 char * 1066 mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp) 1067 { 1068 1069 if (nbp->nb_len > sizeof(dp->d_name) - 1) { 1070 mbnambuf_init(nbp); 1071 return (NULL); 1072 } 1073 memcpy(dp->d_name, &nbp->nb_buf[0], nbp->nb_len); 1074 dp->d_name[nbp->nb_len] = '\0'; 1075 dp->d_namlen = nbp->nb_len; 1076 1077 mbnambuf_init(nbp); 1078 return (dp->d_name); 1079 } 1080