msdosfs_conv.c revision 228796
1/* $FreeBSD: head/sys/fs/msdosfs/msdosfs_conv.c 228796 2011-12-22 06:31:29Z kevlo $ */ 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 61extern struct iconv_functions *msdosfs_iconv; 62 63static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle); 64static u_char * dos2unixchr(u_char *, const u_char **, size_t *, int, struct msdosfsmount *); 65static u_int16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *); 66static u_char * win2unixchr(u_char *, u_int16_t, struct msdosfsmount *); 67static 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 */ 76static u_char 77unix2dos[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 113static u_char 114dos2unix[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 150static u_char 151u2l[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 187static u_char 188l2u[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 */ 236int 237dos2unixfn(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_char *c, tmpbuf[5]; 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(tmpbuf, (const u_char **)&dn, &i, 261 lower & LCASE_BASE, pmp); 262 while (*c != '\0') { 263 *un++ = *c++; 264 thislong++; 265 } 266 } 267 dn += i; 268 269 /* 270 * Now, if there is an extension then put in a period and copy in 271 * the extension. 272 */ 273 if (*dn != ' ') { 274 *un++ = '.'; 275 thislong++; 276 for (i = 3; i > 0 && *dn != ' ';) { 277 c = dos2unixchr(tmpbuf, (const u_char **)&dn, &i, 278 lower & LCASE_EXT, pmp); 279 while (*c != '\0') { 280 *un++ = *c++; 281 thislong++; 282 } 283 } 284 } 285 *un++ = 0; 286 287 return (thislong); 288} 289 290/* 291 * Convert a unix filename to a DOS filename according to Win95 rules. 292 * If applicable and gen is not 0, it is inserted into the converted 293 * filename as a generation number. 294 * Returns 295 * 0 if name couldn't be converted 296 * 1 if the converted name is the same as the original 297 * (no long filename entry necessary for Win95) 298 * 2 if conversion was successful 299 * 3 if conversion was successful and generation number was inserted 300 */ 301int 302unix2dosfn(un, dn, unlen, gen, pmp) 303 const u_char *un; 304 u_char dn[12]; 305 size_t unlen; 306 u_int gen; 307 struct msdosfsmount *pmp; 308{ 309 ssize_t i, j; 310 int l; 311 int conv = 1; 312 const u_char *cp, *dp, *dp1; 313 u_char gentext[6], *wcp; 314 u_int16_t c; 315 316 /* 317 * Fill the dos filename string with blanks. These are DOS's pad 318 * characters. 319 */ 320 for (i = 0; i < 11; i++) 321 dn[i] = ' '; 322 dn[11] = 0; 323 324 /* 325 * The filenames "." and ".." are handled specially, since they 326 * don't follow dos filename rules. 327 */ 328 if (un[0] == '.' && unlen == 1) { 329 dn[0] = '.'; 330 return gen <= 1; 331 } 332 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 333 dn[0] = '.'; 334 dn[1] = '.'; 335 return gen <= 1; 336 } 337 338 /* 339 * Filenames with only blanks and dots are not allowed! 340 */ 341 for (cp = un, i = unlen; --i >= 0; cp++) 342 if (*cp != ' ' && *cp != '.') 343 break; 344 if (i < 0) 345 return 0; 346 347 348 /* 349 * Filenames with some characters are not allowed! 350 */ 351 for (cp = un, i = unlen; i > 0;) 352 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0) 353 return 0; 354 355 /* 356 * Now find the extension 357 * Note: dot as first char doesn't start extension 358 * and trailing dots and blanks are ignored 359 * Note(2003/7): It seems recent Windows has 360 * defferent rule than this code, that Windows 361 * ignores all dots before extension, and use all 362 * chars as filename except for dots. 363 */ 364 dp = dp1 = 0; 365 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 366 switch (*cp++) { 367 case '.': 368 if (!dp1) 369 dp1 = cp; 370 break; 371 case ' ': 372 break; 373 default: 374 if (dp1) 375 dp = dp1; 376 dp1 = 0; 377 break; 378 } 379 } 380 381 /* 382 * Now convert it (this part is for extension). 383 * As Windows XP do, if it's not ascii char, 384 * this function should return 2 or 3, so that checkng out Unicode name. 385 */ 386 if (dp) { 387 if (dp1) 388 l = dp1 - dp; 389 else 390 l = unlen - (dp - un); 391 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) { 392 c = unix2doschr(&cp, (size_t *)&i, pmp); 393 if (c & 0xff00) { 394 dn[j] = c >> 8; 395 if (++j < 11) { 396 dn[j] = c; 397 if (conv != 3) 398 conv = 2; 399 continue; 400 } else { 401 conv = 3; 402 dn[j-1] = ' '; 403 break; 404 } 405 } else { 406 dn[j] = c; 407 } 408 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3) 409 conv = 2; 410 if (dn[j] == 1) { 411 conv = 3; 412 dn[j] = '_'; 413 } 414 if (dn[j] == 2) { 415 conv = 3; 416 dn[j--] = ' '; 417 } 418 } 419 if (i > 0) 420 conv = 3; 421 dp--; 422 } else { 423 for (dp = cp; *--dp == ' ' || *dp == '.';); 424 dp++; 425 } 426 427 /* 428 * Now convert the rest of the name 429 */ 430 for (i = dp - un, j = 0; un < dp && j < 8; j++) { 431 c = unix2doschr(&un, &i, pmp); 432 if (c & 0xff00) { 433 dn[j] = c >> 8; 434 if (++j < 8) { 435 dn[j] = c; 436 if (conv != 3) 437 conv = 2; 438 continue; 439 } else { 440 conv = 3; 441 dn[j-1] = ' '; 442 break; 443 } 444 } else { 445 dn[j] = c; 446 } 447 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3) 448 conv = 2; 449 if (dn[j] == 1) { 450 conv = 3; 451 dn[j] = '_'; 452 } 453 if (dn[j] == 2) { 454 conv = 3; 455 dn[j--] = ' '; 456 } 457 } 458 if (un < dp) 459 conv = 3; 460 /* 461 * If we didn't have any chars in filename, 462 * generate a default 463 */ 464 if (!j) 465 dn[0] = '_'; 466 467 /* 468 * If there wasn't any char dropped, 469 * there is no place for generation numbers 470 */ 471 if (conv != 3) { 472 if (gen > 1) 473 conv = 0; 474 goto done; 475 } 476 477 /* 478 * Now insert the generation number into the filename part 479 */ 480 if (gen == 0) 481 goto done; 482 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 483 *--wcp = gen % 10 + '0'; 484 if (gen) { 485 conv = 0; 486 goto done; 487 } 488 for (i = 8; dn[--i] == ' ';); 489 i++; 490 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 491 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 492 /* 493 * Correct posision to where insert the generation number 494 */ 495 cp = dn; 496 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u); 497 498 dn[i++] = '~'; 499 while (wcp < gentext + sizeof(gentext)) 500 dn[i++] = *wcp++; 501 502 /* 503 * Tail of the filename should be space 504 */ 505 while (i < 8) 506 dn[i++] = ' '; 507 conv = 3; 508 509done: 510 /* 511 * The first character cannot be E5, 512 * because that means a deleted entry 513 */ 514 if (dn[0] == 0xe5) 515 dn[0] = SLOT_E5; 516 517 return conv; 518} 519 520/* 521 * Create a Win95 long name directory entry 522 * Note: assumes that the filename is valid, 523 * i.e. doesn't consist solely of blanks and dots 524 */ 525int 526unix2winfn(un, unlen, wep, cnt, chksum, pmp) 527 const u_char *un; 528 size_t unlen; 529 struct winentry *wep; 530 int cnt; 531 int chksum; 532 struct msdosfsmount *pmp; 533{ 534 u_int8_t *wcp; 535 int i, end; 536 u_int16_t code; 537 538 /* 539 * Drop trailing blanks and dots 540 */ 541 unlen = winLenFixup(un, unlen); 542 543 /* 544 * Cut *un for this slot 545 */ 546 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2, 547 pmp->pm_flags, pmp->pm_u2w); 548 549 /* 550 * Initialize winentry to some useful default 551 */ 552 for (wcp = (u_int8_t *)wep, i = sizeof(*wep); --i >= 0; *wcp++ = 0xff); 553 wep->weCnt = cnt; 554 wep->weAttributes = ATTR_WIN95; 555 wep->weReserved1 = 0; 556 wep->weChksum = chksum; 557 wep->weReserved2 = 0; 558 559 /* 560 * Now convert the filename parts 561 */ 562 end = 0; 563 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) { 564 code = unix2winchr(&un, &unlen, 0, pmp); 565 *wcp++ = code; 566 *wcp++ = code >> 8; 567 if (!code) 568 end = WIN_LAST; 569 } 570 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) { 571 code = unix2winchr(&un, &unlen, 0, pmp); 572 *wcp++ = code; 573 *wcp++ = code >> 8; 574 if (!code) 575 end = WIN_LAST; 576 } 577 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) { 578 code = unix2winchr(&un, &unlen, 0, pmp); 579 *wcp++ = code; 580 *wcp++ = code >> 8; 581 if (!code) 582 end = WIN_LAST; 583 } 584 if (*un == '\0') 585 end = WIN_LAST; 586 wep->weCnt |= end; 587 return !end; 588} 589 590/* 591 * Compare our filename to the one in the Win95 entry 592 * Returns the checksum or -1 if no match 593 */ 594int 595winChkName(nbp, un, unlen, chksum, pmp) 596 struct mbnambuf *nbp; 597 const u_char *un; 598 size_t unlen; 599 int chksum; 600 struct msdosfsmount *pmp; 601{ 602 size_t len; 603 u_int16_t c1, c2; 604 u_char *np; 605 struct dirent dirbuf; 606 607 /* 608 * We already have winentry in *nbp. 609 */ 610 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0) 611 return -1; 612 613#ifdef MSDOSFS_DEBUG 614 printf("winChkName(): un=%s:%d,d_name=%s:%d\n", un, unlen, 615 dirbuf.d_name, 616 dirbuf.d_namlen); 617#endif 618 619 /* 620 * Compare the name parts 621 */ 622 len = dirbuf.d_namlen; 623 if (unlen != len) 624 return -2; 625 626 for (np = dirbuf.d_name; unlen > 0 && len > 0;) { 627 /* 628 * Comparison must be case insensitive, because FAT disallows 629 * to look up or create files in case sensitive even when 630 * it's a long file name. 631 */ 632 c1 = unix2winchr((const u_char **)&np, &len, LCASE_BASE, pmp); 633 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp); 634 if (c1 != c2) 635 return -2; 636 } 637 return chksum; 638} 639 640/* 641 * Convert Win95 filename to dirbuf. 642 * Returns the checksum or -1 if impossible 643 */ 644int 645win2unixfn(nbp, wep, chksum, pmp) 646 struct mbnambuf *nbp; 647 struct winentry *wep; 648 int chksum; 649 struct msdosfsmount *pmp; 650{ 651 u_char *c, tmpbuf[5]; 652 u_int8_t *cp; 653 u_int8_t *np, name[WIN_CHARS * 3 + 1]; 654 u_int16_t code; 655 int i; 656 657 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 658 || !(wep->weCnt&WIN_CNT)) 659 return -1; 660 661 /* 662 * First compare checksums 663 */ 664 if (wep->weCnt&WIN_LAST) { 665 chksum = wep->weChksum; 666 } else if (chksum != wep->weChksum) 667 chksum = -1; 668 if (chksum == -1) 669 return -1; 670 671 /* 672 * Convert the name parts 673 */ 674 np = name; 675 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) { 676 code = (cp[1] << 8) | cp[0]; 677 switch (code) { 678 case 0: 679 *np = '\0'; 680 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 681 return chksum; 682 case '/': 683 *np = '\0'; 684 return -1; 685 default: 686 c = win2unixchr(tmpbuf, code, pmp); 687 while (*c != '\0') 688 *np++ = *c++; 689 break; 690 } 691 cp += 2; 692 } 693 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) { 694 code = (cp[1] << 8) | cp[0]; 695 switch (code) { 696 case 0: 697 *np = '\0'; 698 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 699 return chksum; 700 case '/': 701 *np = '\0'; 702 return -1; 703 default: 704 c = win2unixchr(tmpbuf, code, pmp); 705 while (*c != '\0') 706 *np++ = *c++; 707 break; 708 } 709 cp += 2; 710 } 711 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) { 712 code = (cp[1] << 8) | cp[0]; 713 switch (code) { 714 case 0: 715 *np = '\0'; 716 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 717 return chksum; 718 case '/': 719 *np = '\0'; 720 return -1; 721 default: 722 c = win2unixchr(tmpbuf, code, pmp); 723 while (*c != '\0') 724 *np++ = *c++; 725 break; 726 } 727 cp += 2; 728 } 729 *np = '\0'; 730 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 731 return chksum; 732} 733 734/* 735 * Compute the unrolled checksum of a DOS filename for Win95 LFN use. 736 */ 737u_int8_t 738winChksum(u_int8_t *name) 739{ 740 int i; 741 u_int8_t s; 742 743 for (s = 0, i = 11; --i >= 0; s += *name++) 744 s = (s << 7)|(s >> 1); 745 return (s); 746} 747 748/* 749 * Determine the number of slots necessary for Win95 names 750 */ 751int 752winSlotCnt(un, unlen, pmp) 753 const u_char *un; 754 size_t unlen; 755 struct msdosfsmount *pmp; 756{ 757 size_t wlen; 758 char wn[WIN_MAXLEN * 2 + 1], *wnp; 759 760 unlen = winLenFixup(un, unlen); 761 762 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 763 wlen = WIN_MAXLEN * 2; 764 wnp = wn; 765 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen); 766 if (unlen > 0) 767 return 0; 768 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS); 769 } 770 771 if (unlen > WIN_MAXLEN) 772 return 0; 773 return howmany(unlen, WIN_CHARS); 774} 775 776/* 777 * Determine the number of bytes neccesary for Win95 names 778 */ 779size_t 780winLenFixup(un, unlen) 781 const u_char* un; 782 size_t unlen; 783{ 784 for (un += unlen; unlen > 0; unlen--) 785 if (*--un != ' ' && *un != '.') 786 break; 787 return unlen; 788} 789 790/* 791 * Store an area with multi byte string instr, and reterns left 792 * byte of instr and moves pointer forward. The area's size is 793 * inlen or outlen. 794 */ 795static int 796mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle) 797{ 798 char *outp, outstr[outlen * weight + 1]; 799 800 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) { 801 outp = outstr; 802 outlen *= weight; 803 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen); 804 return (inlen); 805 } 806 807 (*instr) += min(inlen, outlen); 808 return (inlen - min(inlen, outlen)); 809} 810 811/* 812 * Convert DOS char to Local char 813 */ 814static u_char * 815dos2unixchr(u_char *outbuf, const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 816{ 817 u_char c, *outp; 818 size_t len, olen; 819 820 outp = outbuf; 821 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 822 olen = len = 4; 823 824 if (lower & (LCASE_BASE | LCASE_EXT)) 825 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr, 826 ilen, (char **)&outp, &olen, KICONV_LOWER); 827 else 828 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr, 829 ilen, (char **)&outp, &olen); 830 len -= olen; 831 832 /* 833 * return '?' if failed to convert 834 */ 835 if (len == 0) { 836 (*ilen)--; 837 (*instr)++; 838 *outp++ = '?'; 839 } 840 } else { 841 (*ilen)--; 842 c = *(*instr)++; 843 c = dos2unix[c]; 844 if (lower & (LCASE_BASE | LCASE_EXT)) 845 c = u2l[c]; 846 *outp++ = c; 847 outbuf[1] = '\0'; 848 } 849 850 *outp = '\0'; 851 outp = outbuf; 852 return (outp); 853} 854 855/* 856 * Convert Local char to DOS char 857 */ 858static u_int16_t 859unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp) 860{ 861 u_char c; 862 char *up, *outp, unicode[3], outbuf[3]; 863 u_int16_t wc; 864 size_t len, ucslen, unixlen, olen; 865 866 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 867 /* 868 * to hide an invisible character, using a unicode filter 869 */ 870 ucslen = 2; 871 len = *ilen; 872 up = unicode; 873 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 874 ilen, &up, &ucslen); 875 unixlen = len - *ilen; 876 877 /* 878 * cannot be converted 879 */ 880 if (unixlen == 0) { 881 (*ilen)--; 882 (*instr)++; 883 return (0); 884 } 885 886 /* 887 * return magic number for ascii char 888 */ 889 if (unixlen == 1) { 890 c = *(*instr -1); 891 if (! (c & 0x80)) { 892 c = unix2dos[c]; 893 if (c <= 2) 894 return (c); 895 } 896 } 897 898 /* 899 * now convert using libiconv 900 */ 901 *instr -= unixlen; 902 *ilen = len; 903 904 olen = len = 2; 905 outp = outbuf; 906 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr, 907 ilen, &outp, &olen, KICONV_FROM_UPPER); 908 len -= olen; 909 910 /* 911 * cannot be converted, but has unicode char, should return magic number 912 */ 913 if (len == 0) { 914 (*ilen) -= unixlen; 915 (*instr) += unixlen; 916 return (1); 917 } 918 919 wc = 0; 920 while(len--) 921 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 922 return (wc); 923 } 924 925 (*ilen)--; 926 c = *(*instr)++; 927 c = l2u[c]; 928 c = unix2dos[c]; 929 return ((u_int16_t)c); 930} 931 932/* 933 * Convert Windows char to Local char 934 */ 935static u_char * 936win2unixchr(u_char *outbuf, u_int16_t wc, struct msdosfsmount *pmp) 937{ 938 u_char *inp, *outp, inbuf[3]; 939 size_t ilen, olen, len; 940 941 outp = outbuf; 942 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 943 inbuf[0] = (u_char)(wc>>8); 944 inbuf[1] = (u_char)wc; 945 inbuf[2] = '\0'; 946 947 ilen = 2; 948 olen = len = 4; 949 inp = inbuf; 950 msdosfs_iconv->convchr(pmp->pm_w2u, (const char **)&inp, &ilen, 951 (char **)&outp, &olen); 952 len -= olen; 953 954 /* 955 * return '?' if failed to convert 956 */ 957 if (len == 0) 958 *outp++ = '?'; 959 } else { 960 *outp++ = (wc & 0xff00) ? '?' : (u_char)(wc & 0xff); 961 } 962 963 *outp = '\0'; 964 outp = outbuf; 965 return (outp); 966} 967 968/* 969 * Convert Local char to Windows char 970 */ 971static u_int16_t 972unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 973{ 974 u_char *outp, outbuf[3]; 975 u_int16_t wc; 976 size_t olen; 977 978 if (*ilen == 0) 979 return (0); 980 981 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 982 outp = outbuf; 983 olen = 2; 984 if (lower & (LCASE_BASE | LCASE_EXT)) 985 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr, 986 ilen, (char **)&outp, &olen, 987 KICONV_FROM_LOWER); 988 else 989 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 990 ilen, (char **)&outp, &olen); 991 992 /* 993 * return '0' if end of filename 994 */ 995 if (olen == 2) 996 return (0); 997 998 wc = (outbuf[0]<<8) | outbuf[1]; 999 1000 return (wc); 1001 } 1002 1003 (*ilen)--; 1004 wc = (*instr)[0]; 1005 if (lower & (LCASE_BASE | LCASE_EXT)) 1006 wc = u2l[wc]; 1007 (*instr)++; 1008 return (wc); 1009} 1010 1011/* 1012 * Initialize the temporary concatenation buffer. 1013 */ 1014void 1015mbnambuf_init(struct mbnambuf *nbp) 1016{ 1017 1018 nbp->nb_len = 0; 1019 nbp->nb_last_id = -1; 1020 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0'; 1021} 1022 1023/* 1024 * Fill out our concatenation buffer with the given substring, at the offset 1025 * specified by its id. Since this function must be called with ids in 1026 * descending order, we take advantage of the fact that ASCII substrings are 1027 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all 1028 * previous (i.e. higher id) substrings upwards to make room for this one. 1029 * This only penalizes portions of substrings that contain more than 1030 * WIN_CHARS bytes when they are first encountered. 1031 */ 1032void 1033mbnambuf_write(struct mbnambuf *nbp, char *name, int id) 1034{ 1035 char *slot; 1036 size_t count, newlen; 1037 1038 if (nbp->nb_len != 0 && id != nbp->nb_last_id - 1) { 1039#ifdef MSDOSFS_DEBUG 1040 printf("msdosfs: non-decreasing id: id %d, last id %d\n", 1041 id, nbp->nb_last_id); 1042#endif 1043 return; 1044 } 1045 1046 /* Will store this substring in a WIN_CHARS-aligned slot. */ 1047 slot = &nbp->nb_buf[id * WIN_CHARS]; 1048 count = strlen(name); 1049 newlen = nbp->nb_len + count; 1050 if (newlen > WIN_MAXLEN || newlen > MAXNAMLEN) { 1051#ifdef MSDOSFS_DEBUG 1052 printf("msdosfs: file name length %zu too large\n", newlen); 1053#endif 1054 return; 1055 } 1056 1057 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */ 1058 if (count > WIN_CHARS && nbp->nb_len != 0) 1059 bcopy(slot + WIN_CHARS, slot + count, nbp->nb_len); 1060 1061 /* Copy in the substring to its slot and update length so far. */ 1062 bcopy(name, slot, count); 1063 nbp->nb_len = newlen; 1064 nbp->nb_last_id = id; 1065} 1066 1067/* 1068 * Take the completed string and use it to setup the struct dirent. 1069 * Be sure to always nul-terminate the d_name and then copy the string 1070 * from our buffer. Note that this function assumes the full string has 1071 * been reassembled in the buffer. If it's called before all substrings 1072 * have been written via mbnambuf_write(), the result will be incorrect. 1073 */ 1074char * 1075mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp) 1076{ 1077 1078 if (nbp->nb_len > sizeof(dp->d_name) - 1) { 1079 mbnambuf_init(nbp); 1080 return (NULL); 1081 } 1082 bcopy(&nbp->nb_buf[0], dp->d_name, nbp->nb_len); 1083 dp->d_name[nbp->nb_len] = '\0'; 1084 dp->d_namlen = nbp->nb_len; 1085 1086 mbnambuf_init(nbp); 1087 return (dp->d_name); 1088} 1089