1 /* $FreeBSD$ */ 2 /* $NetBSD: msdosfs_conv.c,v 1.25 1997/11/17 15:36:40 ws Exp $ */ 3 4 /*- 5 * Copyright (C) 1995, 1997 Wolfgang Solfrank. 6 * Copyright (C) 1995, 1997 TooLs GmbH. 7 * All rights reserved. 8 * Original code by Paul Popelka (paulp@uts.amdahl.com) (see below). 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. All advertising materials mentioning features or use of this software 19 * must display the following acknowledgement: 20 * This product includes software developed by TooLs GmbH. 21 * 4. The name of TooLs GmbH may not be used to endorse or promote products 22 * derived from this software without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR 25 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 26 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 27 * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 28 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 29 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; 30 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, 31 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR 32 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF 33 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 34 */ 35 /*- 36 * Written by Paul Popelka (paulp@uts.amdahl.com) 37 * 38 * You can do anything you want with this software, just don't say you wrote 39 * it, and don't remove this notice. 40 * 41 * This software is provided "as is". 42 * 43 * The author supplies this software to be publicly redistributed on the 44 * understanding that the author is not responsible for the correct 45 * functioning of this software in any circumstances and is not liable for 46 * any damages caused by this software. 47 * 48 * October 1992 49 */ 50 51 #include <sys/param.h> 52 #include <sys/systm.h> 53 #include <sys/dirent.h> 54 #include <sys/iconv.h> 55 #include <sys/mount.h> 56 57 #include <fs/msdosfs/bpb.h> 58 #include <fs/msdosfs/direntry.h> 59 #include <fs/msdosfs/msdosfsmount.h> 60 61 extern struct iconv_functions *msdosfs_iconv; 62 63 static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle); 64 static u_int16_t dos2unixchr(const u_char **, size_t *, int, struct msdosfsmount *); 65 static u_int16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *); 66 static u_int16_t win2unixchr(u_int16_t, struct msdosfsmount *); 67 static u_int16_t unix2winchr(const u_char **, size_t *, int, struct msdosfsmount *); 68 69 /* 70 * 0 - character disallowed in long file name. 71 * 1 - character should be replaced by '_' in DOS file name, 72 * and generation number inserted. 73 * 2 - character ('.' and ' ') should be skipped in DOS file name, 74 * and generation number inserted. 75 */ 76 static u_char 77 unix2dos[256] = { 78 /* iso8859-1 -> cp850 */ 79 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */ 80 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */ 81 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */ 82 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */ 83 2, 0x21, 0, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 84 0x28, 0x29, 0, 1, 1, 0x2d, 2, 0, /* 28-2f */ 85 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 86 0x38, 0x39, 0, 1, 0, 1, 0, 0, /* 38-3f */ 87 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 88 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 89 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 90 0x58, 0x59, 0x5a, 1, 0, 1, 0x5e, 0x5f, /* 58-5f */ 91 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 60-67 */ 92 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 68-6f */ 93 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 70-77 */ 94 0x58, 0x59, 0x5a, 0x7b, 0, 0x7d, 0x7e, 0, /* 78-7f */ 95 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */ 96 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */ 97 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */ 98 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */ 99 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */ 100 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */ 101 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */ 102 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */ 103 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */ 104 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */ 105 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */ 106 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */ 107 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */ 108 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */ 109 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */ 110 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */ 111 }; 112 113 static u_char 114 dos2unix[256] = { 115 /* cp850 -> iso8859-1 */ 116 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 00-07 */ 117 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 08-0f */ 118 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 10-17 */ 119 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 18-1f */ 120 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 121 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 122 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 123 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 124 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */ 125 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */ 126 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */ 127 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 128 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 129 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 130 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 131 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 132 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */ 133 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */ 134 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */ 135 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, 0x3f, /* 98-9f */ 136 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */ 137 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */ 138 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xc1, 0xc2, 0xc0, /* b0-b7 */ 139 0xa9, 0x3f, 0x3f, 0x3f, 0x3f, 0xa2, 0xa5, 0x3f, /* b8-bf */ 140 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xe3, 0xc3, /* c0-c7 */ 141 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xa4, /* c8-cf */ 142 0xf0, 0xd0, 0xca, 0xcb, 0xc8, 0x3f, 0xcd, 0xce, /* d0-d7 */ 143 0xcf, 0x3f, 0x3f, 0x3f, 0x3f, 0xa6, 0xcc, 0x3f, /* d8-df */ 144 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */ 145 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */ 146 0xad, 0xb1, 0x3f, 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */ 147 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, 0x3f, 0x3f, /* f8-ff */ 148 }; 149 150 static u_char 151 u2l[256] = { 152 /* tolower */ 153 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 154 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 155 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 156 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 157 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 158 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 159 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 160 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 161 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 162 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 163 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 164 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 165 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 166 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 167 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 168 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 169 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 170 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 171 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 172 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 173 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 174 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 175 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 176 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 177 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 178 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 179 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 180 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 181 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 182 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 183 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 184 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 185 }; 186 187 static u_char 188 l2u[256] = { 189 /* toupper */ 190 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */ 191 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */ 192 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */ 193 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */ 194 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */ 195 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */ 196 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */ 197 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */ 198 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */ 199 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */ 200 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */ 201 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */ 202 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */ 203 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */ 204 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */ 205 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */ 206 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */ 207 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */ 208 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */ 209 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */ 210 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */ 211 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */ 212 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */ 213 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */ 214 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */ 215 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */ 216 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */ 217 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */ 218 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */ 219 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */ 220 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */ 221 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */ 222 }; 223 224 /* 225 * DOS filenames are made of 2 parts, the name part and the extension part. 226 * The name part is 8 characters long and the extension part is 3 227 * characters long. They may contain trailing blanks if the name or 228 * extension are not long enough to fill their respective fields. 229 */ 230 231 /* 232 * Convert a DOS filename to a unix filename. And, return the number of 233 * characters in the resulting unix filename excluding the terminating 234 * null. 235 */ 236 int 237 dos2unixfn(dn, un, lower, pmp) 238 u_char dn[11]; 239 u_char *un; 240 int lower; 241 struct msdosfsmount *pmp; 242 { 243 size_t i; 244 int thislong = 0; 245 u_int16_t c; 246 247 /* 248 * If first char of the filename is SLOT_E5 (0x05), then the real 249 * first char of the filename should be 0xe5. But, they couldn't 250 * just have a 0xe5 mean 0xe5 because that is used to mean a freed 251 * directory slot. Another dos quirk. 252 */ 253 if (*dn == SLOT_E5) 254 *dn = 0xe5; 255 256 /* 257 * Copy the name portion into the unix filename string. 258 */ 259 for (i = 8; i > 0 && *dn != ' ';) { 260 c = dos2unixchr((const u_char **)&dn, &i, lower & LCASE_BASE, 261 pmp); 262 if (c & 0xff00) { 263 *un++ = c >> 8; 264 thislong++; 265 } 266 *un++ = c; 267 thislong++; 268 } 269 dn += i; 270 271 /* 272 * Now, if there is an extension then put in a period and copy in 273 * the extension. 274 */ 275 if (*dn != ' ') { 276 *un++ = '.'; 277 thislong++; 278 for (i = 3; i > 0 && *dn != ' ';) { 279 c = dos2unixchr((const u_char **)&dn, &i, 280 lower & LCASE_EXT, pmp); 281 if (c & 0xff00) { 282 *un++ = c >> 8; 283 thislong++; 284 } 285 *un++ = c; 286 thislong++; 287 } 288 } 289 *un++ = 0; 290 291 return (thislong); 292 } 293 294 /* 295 * Convert a unix filename to a DOS filename according to Win95 rules. 296 * If applicable and gen is not 0, it is inserted into the converted 297 * filename as a generation number. 298 * Returns 299 * 0 if name couldn't be converted 300 * 1 if the converted name is the same as the original 301 * (no long filename entry necessary for Win95) 302 * 2 if conversion was successful 303 * 3 if conversion was successful and generation number was inserted 304 */ 305 int 306 unix2dosfn(un, dn, unlen, gen, pmp) 307 const u_char *un; 308 u_char dn[12]; 309 size_t unlen; 310 u_int gen; 311 struct msdosfsmount *pmp; 312 { 313 ssize_t i, j; 314 int l; 315 int conv = 1; 316 const u_char *cp, *dp, *dp1; 317 u_char gentext[6], *wcp; 318 u_int16_t c; 319 320 /* 321 * Fill the dos filename string with blanks. These are DOS's pad 322 * characters. 323 */ 324 for (i = 0; i < 11; i++) 325 dn[i] = ' '; 326 dn[11] = 0; 327 328 /* 329 * The filenames "." and ".." are handled specially, since they 330 * don't follow dos filename rules. 331 */ 332 if (un[0] == '.' && unlen == 1) { 333 dn[0] = '.'; 334 return gen <= 1; 335 } 336 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 337 dn[0] = '.'; 338 dn[1] = '.'; 339 return gen <= 1; 340 } 341 342 /* 343 * Filenames with only blanks and dots are not allowed! 344 */ 345 for (cp = un, i = unlen; --i >= 0; cp++) 346 if (*cp != ' ' && *cp != '.') 347 break; 348 if (i < 0) 349 return 0; 350 351 352 /* 353 * Filenames with some characters are not allowed! 354 */ 355 for (cp = un, i = unlen; i > 0;) 356 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0) 357 return 0; 358 359 /* 360 * Now find the extension 361 * Note: dot as first char doesn't start extension 362 * and trailing dots and blanks are ignored 363 * Note(2003/7): It seems recent Windows has 364 * defferent rule than this code, that Windows 365 * ignores all dots before extension, and use all 366 * chars as filename except for dots. 367 */ 368 dp = dp1 = 0; 369 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 370 switch (*cp++) { 371 case '.': 372 if (!dp1) 373 dp1 = cp; 374 break; 375 case ' ': 376 break; 377 default: 378 if (dp1) 379 dp = dp1; 380 dp1 = 0; 381 break; 382 } 383 } 384 385 /* 386 * Now convert it (this part is for extension). 387 * As Windows XP do, if it's not ascii char, 388 * this function should return 2 or 3, so that checkng out Unicode name. 389 */ 390 if (dp) { 391 if (dp1) 392 l = dp1 - dp; 393 else 394 l = unlen - (dp - un); 395 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) { 396 c = unix2doschr(&cp, (size_t *)&i, pmp); 397 if (c & 0xff00) { 398 dn[j] = c >> 8; 399 if (++j < 11) { 400 dn[j] = c; 401 if (conv != 3) 402 conv = 2; 403 continue; 404 } else { 405 conv = 3; 406 dn[j-1] = ' '; 407 break; 408 } 409 } else { 410 dn[j] = c; 411 } 412 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3) 413 conv = 2; 414 if (dn[j] == 1) { 415 conv = 3; 416 dn[j] = '_'; 417 } 418 if (dn[j] == 2) { 419 conv = 3; 420 dn[j--] = ' '; 421 } 422 } 423 if (i > 0) 424 conv = 3; 425 dp--; 426 } else { 427 for (dp = cp; *--dp == ' ' || *dp == '.';); 428 dp++; 429 } 430 431 /* 432 * Now convert the rest of the name 433 */ 434 for (i = dp - un, j = 0; un < dp && j < 8; j++) { 435 c = unix2doschr(&un, &i, pmp); 436 if (c & 0xff00) { 437 dn[j] = c >> 8; 438 if (++j < 8) { 439 dn[j] = c; 440 if (conv != 3) 441 conv = 2; 442 continue; 443 } else { 444 conv = 3; 445 dn[j-1] = ' '; 446 break; 447 } 448 } else { 449 dn[j] = c; 450 } 451 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3) 452 conv = 2; 453 if (dn[j] == 1) { 454 conv = 3; 455 dn[j] = '_'; 456 } 457 if (dn[j] == 2) { 458 conv = 3; 459 dn[j--] = ' '; 460 } 461 } 462 if (un < dp) 463 conv = 3; 464 /* 465 * If we didn't have any chars in filename, 466 * generate a default 467 */ 468 if (!j) 469 dn[0] = '_'; 470 471 /* 472 * If there wasn't any char dropped, 473 * there is no place for generation numbers 474 */ 475 if (conv != 3) { 476 if (gen > 1) 477 conv = 0; 478 goto done; 479 } 480 481 /* 482 * Now insert the generation number into the filename part 483 */ 484 if (gen == 0) 485 goto done; 486 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 487 *--wcp = gen % 10 + '0'; 488 if (gen) { 489 conv = 0; 490 goto done; 491 } 492 for (i = 8; dn[--i] == ' ';); 493 i++; 494 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 495 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 496 /* 497 * Correct posision to where insert the generation number 498 */ 499 cp = dn; 500 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u); 501 502 dn[i++] = '~'; 503 while (wcp < gentext + sizeof(gentext)) 504 dn[i++] = *wcp++; 505 506 /* 507 * Tail of the filename should be space 508 */ 509 while (i < 8) 510 dn[i++] = ' '; 511 conv = 3; 512 513 done: 514 /* 515 * The first character cannot be E5, 516 * because that means a deleted entry 517 */ 518 if (dn[0] == 0xe5) 519 dn[0] = SLOT_E5; 520 521 return conv; 522 } 523 524 /* 525 * Create a Win95 long name directory entry 526 * Note: assumes that the filename is valid, 527 * i.e. doesn't consist solely of blanks and dots 528 */ 529 int 530 unix2winfn(un, unlen, wep, cnt, chksum, pmp) 531 const u_char *un; 532 size_t unlen; 533 struct winentry *wep; 534 int cnt; 535 int chksum; 536 struct msdosfsmount *pmp; 537 { 538 u_int8_t *wcp; 539 int i, end; 540 u_int16_t code; 541 542 /* 543 * Drop trailing blanks and dots 544 */ 545 unlen = winLenFixup(un, unlen); 546 547 /* 548 * Cut *un for this slot 549 */ 550 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2, 551 pmp->pm_flags, pmp->pm_u2w); 552 553 /* 554 * Initialize winentry to some useful default 555 */ 556 for (wcp = (u_int8_t *)wep, i = sizeof(*wep); --i >= 0; *wcp++ = 0xff); 557 wep->weCnt = cnt; 558 wep->weAttributes = ATTR_WIN95; 559 wep->weReserved1 = 0; 560 wep->weChksum = chksum; 561 wep->weReserved2 = 0; 562 563 /* 564 * Now convert the filename parts 565 */ 566 end = 0; 567 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) { 568 code = unix2winchr(&un, &unlen, 0, pmp); 569 *wcp++ = code; 570 *wcp++ = code >> 8; 571 if (!code) 572 end = WIN_LAST; 573 } 574 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) { 575 code = unix2winchr(&un, &unlen, 0, pmp); 576 *wcp++ = code; 577 *wcp++ = code >> 8; 578 if (!code) 579 end = WIN_LAST; 580 } 581 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) { 582 code = unix2winchr(&un, &unlen, 0, pmp); 583 *wcp++ = code; 584 *wcp++ = code >> 8; 585 if (!code) 586 end = WIN_LAST; 587 } 588 if (*un == '\0') 589 end = WIN_LAST; 590 wep->weCnt |= end; 591 return !end; 592 } 593 594 /* 595 * Compare our filename to the one in the Win95 entry 596 * Returns the checksum or -1 if no match 597 */ 598 int 599 winChkName(nbp, un, unlen, chksum, pmp) 600 struct mbnambuf *nbp; 601 const u_char *un; 602 size_t unlen; 603 int chksum; 604 struct msdosfsmount *pmp; 605 { 606 size_t len; 607 u_int16_t c1, c2; 608 u_char *np; 609 struct dirent dirbuf; 610 611 /* 612 * We already have winentry in *nbp. 613 */ 614 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0) 615 return -1; 616 617 #ifdef MSDOSFS_DEBUG 618 printf("winChkName(): un=%s:%d,d_name=%s:%d\n", un, unlen, 619 dirbuf.d_name, 620 dirbuf.d_namlen); 621 #endif 622 623 /* 624 * Compare the name parts 625 */ 626 len = dirbuf.d_namlen; 627 if (unlen != len) 628 return -2; 629 630 for (np = dirbuf.d_name; unlen > 0 && len > 0;) { 631 /* 632 * Comparison must be case insensitive, because FAT disallows 633 * to look up or create files in case sensitive even when 634 * it's a long file name. 635 */ 636 c1 = unix2winchr((const u_char **)&np, &len, LCASE_BASE, pmp); 637 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp); 638 if (c1 != c2) 639 return -2; 640 } 641 return chksum; 642 } 643 644 /* 645 * Convert Win95 filename to dirbuf. 646 * Returns the checksum or -1 if impossible 647 */ 648 int 649 win2unixfn(nbp, wep, chksum, pmp) 650 struct mbnambuf *nbp; 651 struct winentry *wep; 652 int chksum; 653 struct msdosfsmount *pmp; 654 { 655 u_int8_t *cp; 656 u_int8_t *np, name[WIN_CHARS * 2 + 1]; 657 u_int16_t code; 658 int i; 659 660 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 661 || !(wep->weCnt&WIN_CNT)) 662 return -1; 663 664 /* 665 * First compare checksums 666 */ 667 if (wep->weCnt&WIN_LAST) { 668 chksum = wep->weChksum; 669 } else if (chksum != wep->weChksum) 670 chksum = -1; 671 if (chksum == -1) 672 return -1; 673 674 /* 675 * Convert the name parts 676 */ 677 np = name; 678 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) { 679 code = (cp[1] << 8) | cp[0]; 680 switch (code) { 681 case 0: 682 *np = '\0'; 683 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 684 return chksum; 685 case '/': 686 *np = '\0'; 687 return -1; 688 default: 689 code = win2unixchr(code, pmp); 690 if (code & 0xff00) 691 *np++ = code >> 8; 692 *np++ = code; 693 break; 694 } 695 cp += 2; 696 } 697 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) { 698 code = (cp[1] << 8) | cp[0]; 699 switch (code) { 700 case 0: 701 *np = '\0'; 702 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 703 return chksum; 704 case '/': 705 *np = '\0'; 706 return -1; 707 default: 708 code = win2unixchr(code, pmp); 709 if (code & 0xff00) 710 *np++ = code >> 8; 711 *np++ = code; 712 break; 713 } 714 cp += 2; 715 } 716 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) { 717 code = (cp[1] << 8) | cp[0]; 718 switch (code) { 719 case 0: 720 *np = '\0'; 721 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 722 return chksum; 723 case '/': 724 *np = '\0'; 725 return -1; 726 default: 727 code = win2unixchr(code, pmp); 728 if (code & 0xff00) 729 *np++ = code >> 8; 730 *np++ = code; 731 break; 732 } 733 cp += 2; 734 } 735 *np = '\0'; 736 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 737 return chksum; 738 } 739 740 /* 741 * Compute the unrolled checksum of a DOS filename for Win95 LFN use. 742 */ 743 u_int8_t 744 winChksum(struct direntry *dep) 745 { 746 u_int8_t s; 747 748 s = dep->deName[0]; 749 s = ((s << 7) | (s >> 1)) + dep->deName[1]; 750 s = ((s << 7) | (s >> 1)) + dep->deName[2]; 751 s = ((s << 7) | (s >> 1)) + dep->deName[3]; 752 s = ((s << 7) | (s >> 1)) + dep->deName[4]; 753 s = ((s << 7) | (s >> 1)) + dep->deName[5]; 754 s = ((s << 7) | (s >> 1)) + dep->deName[6]; 755 s = ((s << 7) | (s >> 1)) + dep->deName[7]; 756 s = ((s << 7) | (s >> 1)) + dep->deExtension[0]; 757 s = ((s << 7) | (s >> 1)) + dep->deExtension[1]; 758 s = ((s << 7) | (s >> 1)) + dep->deExtension[2]; 759 760 return (s); 761 } 762 763 /* 764 * Determine the number of slots necessary for Win95 names 765 */ 766 int 767 winSlotCnt(un, unlen, pmp) 768 const u_char *un; 769 size_t unlen; 770 struct msdosfsmount *pmp; 771 { 772 size_t wlen; 773 char wn[WIN_MAXLEN * 2 + 1], *wnp; 774 775 unlen = winLenFixup(un, unlen); 776 777 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 778 wlen = WIN_MAXLEN * 2; 779 wnp = wn; 780 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen); 781 if (unlen > 0) 782 return 0; 783 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS); 784 } 785 786 if (unlen > WIN_MAXLEN) 787 return 0; 788 return howmany(unlen, WIN_CHARS); 789 } 790 791 /* 792 * Determine the number of bytes neccesary for Win95 names 793 */ 794 size_t 795 winLenFixup(un, unlen) 796 const u_char* un; 797 size_t unlen; 798 { 799 for (un += unlen; unlen > 0; unlen--) 800 if (*--un != ' ' && *un != '.') 801 break; 802 return unlen; 803 } 804 805 /* 806 * Store an area with multi byte string instr, and reterns left 807 * byte of instr and moves pointer forward. The area's size is 808 * inlen or outlen. 809 */ 810 static int 811 mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle) 812 { 813 char *outp, outstr[outlen * weight + 1]; 814 815 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) { 816 outp = outstr; 817 outlen *= weight; 818 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen); 819 return (inlen); 820 } 821 822 (*instr) += min(inlen, outlen); 823 return (inlen - min(inlen, outlen)); 824 } 825 826 /* 827 * Convert DOS char to Local char 828 */ 829 static u_int16_t 830 dos2unixchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 831 { 832 u_char c; 833 char *outp, outbuf[3]; 834 u_int16_t wc; 835 size_t len, olen; 836 837 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 838 olen = len = 2; 839 outp = outbuf; 840 841 if (lower & (LCASE_BASE | LCASE_EXT)) 842 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr, 843 ilen, &outp, &olen, KICONV_LOWER); 844 else 845 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr, 846 ilen, &outp, &olen); 847 len -= olen; 848 849 /* 850 * return '?' if failed to convert 851 */ 852 if (len == 0) { 853 (*ilen)--; 854 (*instr)++; 855 return ('?'); 856 } 857 858 wc = 0; 859 while(len--) 860 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 861 return (wc); 862 } 863 864 (*ilen)--; 865 c = *(*instr)++; 866 c = dos2unix[c]; 867 if (lower & (LCASE_BASE | LCASE_EXT)) 868 c = u2l[c]; 869 return ((u_int16_t)c); 870 } 871 872 /* 873 * Convert Local char to DOS char 874 */ 875 static u_int16_t 876 unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp) 877 { 878 u_char c; 879 char *up, *outp, unicode[3], outbuf[3]; 880 u_int16_t wc; 881 size_t len, ucslen, unixlen, olen; 882 883 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 884 /* 885 * to hide an invisible character, using a unicode filter 886 */ 887 ucslen = 2; 888 len = *ilen; 889 up = unicode; 890 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 891 ilen, &up, &ucslen); 892 unixlen = len - *ilen; 893 894 /* 895 * cannot be converted 896 */ 897 if (unixlen == 0) { 898 (*ilen)--; 899 (*instr)++; 900 return (0); 901 } 902 903 /* 904 * return magic number for ascii char 905 */ 906 if (unixlen == 1) { 907 c = *(*instr -1); 908 if (! (c & 0x80)) { 909 c = unix2dos[c]; 910 if (c <= 2) 911 return (c); 912 } 913 } 914 915 /* 916 * now convert using libiconv 917 */ 918 *instr -= unixlen; 919 *ilen = len; 920 921 olen = len = 2; 922 outp = outbuf; 923 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr, 924 ilen, &outp, &olen, KICONV_FROM_UPPER); 925 len -= olen; 926 927 /* 928 * cannot be converted, but has unicode char, should return magic number 929 */ 930 if (len == 0) { 931 (*ilen) -= unixlen; 932 (*instr) += unixlen; 933 return (1); 934 } 935 936 wc = 0; 937 while(len--) 938 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 939 return (wc); 940 } 941 942 (*ilen)--; 943 c = *(*instr)++; 944 c = l2u[c]; 945 c = unix2dos[c]; 946 return ((u_int16_t)c); 947 } 948 949 /* 950 * Convert Windows char to Local char 951 */ 952 static u_int16_t 953 win2unixchr(u_int16_t wc, struct msdosfsmount *pmp) 954 { 955 u_char *inp, *outp, inbuf[3], outbuf[3]; 956 size_t ilen, olen, len; 957 958 if (wc == 0) 959 return (0); 960 961 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 962 inbuf[0] = (u_char)(wc>>8); 963 inbuf[1] = (u_char)wc; 964 inbuf[2] = '\0'; 965 966 ilen = olen = len = 2; 967 inp = inbuf; 968 outp = outbuf; 969 msdosfs_iconv->convchr(pmp->pm_w2u, (const char **)&inp, &ilen, 970 (char **)&outp, &olen); 971 len -= olen; 972 973 /* 974 * return '?' if failed to convert 975 */ 976 if (len == 0) { 977 wc = '?'; 978 return (wc); 979 } 980 981 wc = 0; 982 while(len--) 983 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 984 return (wc); 985 } 986 987 if (wc & 0xff00) 988 wc = '?'; 989 990 return (wc); 991 } 992 993 /* 994 * Convert Local char to Windows char 995 */ 996 static u_int16_t 997 unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 998 { 999 u_char *outp, outbuf[3]; 1000 u_int16_t wc; 1001 size_t olen; 1002 1003 if (*ilen == 0) 1004 return (0); 1005 1006 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 1007 outp = outbuf; 1008 olen = 2; 1009 if (lower & (LCASE_BASE | LCASE_EXT)) 1010 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr, 1011 ilen, (char **)&outp, &olen, 1012 KICONV_FROM_LOWER); 1013 else 1014 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 1015 ilen, (char **)&outp, &olen); 1016 1017 /* 1018 * return '0' if end of filename 1019 */ 1020 if (olen == 2) 1021 return (0); 1022 1023 wc = (outbuf[0]<<8) | outbuf[1]; 1024 1025 return (wc); 1026 } 1027 1028 (*ilen)--; 1029 wc = (*instr)[0]; 1030 if (lower & (LCASE_BASE | LCASE_EXT)) 1031 wc = u2l[wc]; 1032 (*instr)++; 1033 return (wc); 1034 } 1035 1036 /* 1037 * Initialize the temporary concatenation buffer. 1038 */ 1039 void 1040 mbnambuf_init(struct mbnambuf *nbp) 1041 { 1042 1043 nbp->nb_len = 0; 1044 nbp->nb_last_id = -1; 1045 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0'; 1046 } 1047 1048 /* 1049 * Fill out our concatenation buffer with the given substring, at the offset 1050 * specified by its id. Since this function must be called with ids in 1051 * descending order, we take advantage of the fact that ASCII substrings are 1052 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all 1053 * previous (i.e. higher id) substrings upwards to make room for this one. 1054 * This only penalizes portions of substrings that contain more than 1055 * WIN_CHARS bytes when they are first encountered. 1056 */ 1057 void 1058 mbnambuf_write(struct mbnambuf *nbp, char *name, int id) 1059 { 1060 char *slot; 1061 size_t count, newlen; 1062 1063 KASSERT(nbp->nb_len == 0 || id == nbp->nb_last_id - 1, 1064 ("non-decreasing id: id %d, last id %d", id, nbp->nb_last_id)); 1065 1066 /* Will store this substring in a WIN_CHARS-aligned slot. */ 1067 slot = &nbp->nb_buf[id * WIN_CHARS]; 1068 count = strlen(name); 1069 newlen = nbp->nb_len + count; 1070 if (newlen > WIN_MAXLEN || newlen > MAXNAMLEN) { 1071 printf("msdosfs: file name length %zu too large\n", newlen); 1072 return; 1073 } 1074 1075 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */ 1076 if (count > WIN_CHARS && nbp->nb_len != 0) 1077 bcopy(slot + WIN_CHARS, slot + count, nbp->nb_len); 1078 1079 /* Copy in the substring to its slot and update length so far. */ 1080 bcopy(name, slot, count); 1081 nbp->nb_len = newlen; 1082 nbp->nb_last_id = id; 1083 } 1084 1085 /* 1086 * Take the completed string and use it to setup the struct dirent. 1087 * Be sure to always nul-terminate the d_name and then copy the string 1088 * from our buffer. Note that this function assumes the full string has 1089 * been reassembled in the buffer. If it's called before all substrings 1090 * have been written via mbnambuf_write(), the result will be incorrect. 1091 */ 1092 char * 1093 mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp) 1094 { 1095 1096 if (nbp->nb_len > sizeof(dp->d_name) - 1) { 1097 mbnambuf_init(nbp); 1098 return (NULL); 1099 } 1100 bcopy(&nbp->nb_buf[0], dp->d_name, nbp->nb_len); 1101 dp->d_name[nbp->nb_len] = '\0'; 1102 dp->d_namlen = nbp->nb_len; 1103 1104 mbnambuf_init(nbp); 1105 return (dp->d_name); 1106 } 1107