msdosfs_conv.c revision 277092
1/* $FreeBSD: head/sys/fs/msdosfs/msdosfs_conv.c 277092 2015-01-12 21:55:48Z emaste $ */ 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(u_char dn[11], u_char *un, int lower, struct msdosfsmount *pmp) 238{ 239 size_t i; 240 int thislong = 0; 241 u_char *c, tmpbuf[5]; 242 243 /* 244 * If first char of the filename is SLOT_E5 (0x05), then the real 245 * first char of the filename should be 0xe5. But, they couldn't 246 * just have a 0xe5 mean 0xe5 because that is used to mean a freed 247 * directory slot. Another dos quirk. 248 */ 249 if (*dn == SLOT_E5) 250 *dn = 0xe5; 251 252 /* 253 * Copy the name portion into the unix filename string. 254 */ 255 for (i = 8; i > 0 && *dn != ' ';) { 256 c = dos2unixchr(tmpbuf, (const u_char **)&dn, &i, 257 lower & LCASE_BASE, pmp); 258 while (*c != '\0') { 259 *un++ = *c++; 260 thislong++; 261 } 262 } 263 dn += i; 264 265 /* 266 * Now, if there is an extension then put in a period and copy in 267 * the extension. 268 */ 269 if (*dn != ' ') { 270 *un++ = '.'; 271 thislong++; 272 for (i = 3; i > 0 && *dn != ' ';) { 273 c = dos2unixchr(tmpbuf, (const u_char **)&dn, &i, 274 lower & LCASE_EXT, pmp); 275 while (*c != '\0') { 276 *un++ = *c++; 277 thislong++; 278 } 279 } 280 } 281 *un++ = 0; 282 283 return (thislong); 284} 285 286/* 287 * Convert a unix filename to a DOS filename according to Win95 rules. 288 * If applicable and gen is not 0, it is inserted into the converted 289 * filename as a generation number. 290 * Returns 291 * 0 if name couldn't be converted 292 * 1 if the converted name is the same as the original 293 * (no long filename entry necessary for Win95) 294 * 2 if conversion was successful 295 * 3 if conversion was successful and generation number was inserted 296 */ 297int 298unix2dosfn(const u_char *un, u_char dn[12], size_t unlen, u_int gen, 299 struct msdosfsmount *pmp) 300{ 301 ssize_t i, j; 302 int l; 303 int conv = 1; 304 const u_char *cp, *dp, *dp1; 305 u_char gentext[6], *wcp; 306 u_int16_t c; 307 308 /* 309 * Fill the dos filename string with blanks. These are DOS's pad 310 * characters. 311 */ 312 for (i = 0; i < 11; i++) 313 dn[i] = ' '; 314 dn[11] = 0; 315 316 /* 317 * The filenames "." and ".." are handled specially, since they 318 * don't follow dos filename rules. 319 */ 320 if (un[0] == '.' && unlen == 1) { 321 dn[0] = '.'; 322 return gen <= 1; 323 } 324 if (un[0] == '.' && un[1] == '.' && unlen == 2) { 325 dn[0] = '.'; 326 dn[1] = '.'; 327 return gen <= 1; 328 } 329 330 /* 331 * Filenames with only blanks and dots are not allowed! 332 */ 333 for (cp = un, i = unlen; --i >= 0; cp++) 334 if (*cp != ' ' && *cp != '.') 335 break; 336 if (i < 0) 337 return 0; 338 339 340 /* 341 * Filenames with some characters are not allowed! 342 */ 343 for (cp = un, i = unlen; i > 0;) 344 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0) 345 return 0; 346 347 /* 348 * Now find the extension 349 * Note: dot as first char doesn't start extension 350 * and trailing dots and blanks are ignored 351 * Note(2003/7): It seems recent Windows has 352 * defferent rule than this code, that Windows 353 * ignores all dots before extension, and use all 354 * chars as filename except for dots. 355 */ 356 dp = dp1 = 0; 357 for (cp = un + 1, i = unlen - 1; --i >= 0;) { 358 switch (*cp++) { 359 case '.': 360 if (!dp1) 361 dp1 = cp; 362 break; 363 case ' ': 364 break; 365 default: 366 if (dp1) 367 dp = dp1; 368 dp1 = 0; 369 break; 370 } 371 } 372 373 /* 374 * Now convert it (this part is for extension). 375 * As Windows XP do, if it's not ascii char, 376 * this function should return 2 or 3, so that checkng out Unicode name. 377 */ 378 if (dp) { 379 if (dp1) 380 l = dp1 - dp; 381 else 382 l = unlen - (dp - un); 383 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) { 384 c = unix2doschr(&cp, (size_t *)&i, pmp); 385 if (c & 0xff00) { 386 dn[j] = c >> 8; 387 if (++j < 11) { 388 dn[j] = c; 389 if (conv != 3) 390 conv = 2; 391 continue; 392 } else { 393 conv = 3; 394 dn[j-1] = ' '; 395 break; 396 } 397 } else { 398 dn[j] = c; 399 } 400 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3) 401 conv = 2; 402 if (dn[j] == 1) { 403 conv = 3; 404 dn[j] = '_'; 405 } 406 if (dn[j] == 2) { 407 conv = 3; 408 dn[j--] = ' '; 409 } 410 } 411 if (i > 0) 412 conv = 3; 413 dp--; 414 } else { 415 for (dp = cp; *--dp == ' ' || *dp == '.';); 416 dp++; 417 } 418 419 /* 420 * Now convert the rest of the name 421 */ 422 for (i = dp - un, j = 0; un < dp && j < 8; j++) { 423 c = unix2doschr(&un, &i, pmp); 424 if (c & 0xff00) { 425 dn[j] = c >> 8; 426 if (++j < 8) { 427 dn[j] = c; 428 if (conv != 3) 429 conv = 2; 430 continue; 431 } else { 432 conv = 3; 433 dn[j-1] = ' '; 434 break; 435 } 436 } else { 437 dn[j] = c; 438 } 439 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3) 440 conv = 2; 441 if (dn[j] == 1) { 442 conv = 3; 443 dn[j] = '_'; 444 } 445 if (dn[j] == 2) { 446 conv = 3; 447 dn[j--] = ' '; 448 } 449 } 450 if (un < dp) 451 conv = 3; 452 /* 453 * If we didn't have any chars in filename, 454 * generate a default 455 */ 456 if (!j) 457 dn[0] = '_'; 458 459 /* 460 * If there wasn't any char dropped, 461 * there is no place for generation numbers 462 */ 463 if (conv != 3) { 464 if (gen > 1) 465 conv = 0; 466 goto done; 467 } 468 469 /* 470 * Now insert the generation number into the filename part 471 */ 472 if (gen == 0) 473 goto done; 474 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10) 475 *--wcp = gen % 10 + '0'; 476 if (gen) { 477 conv = 0; 478 goto done; 479 } 480 for (i = 8; dn[--i] == ' ';); 481 i++; 482 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i) 483 i = 8 - (gentext + sizeof(gentext) - wcp + 1); 484 /* 485 * Correct posision to where insert the generation number 486 */ 487 cp = dn; 488 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u); 489 490 dn[i++] = '~'; 491 while (wcp < gentext + sizeof(gentext)) 492 dn[i++] = *wcp++; 493 494 /* 495 * Tail of the filename should be space 496 */ 497 while (i < 8) 498 dn[i++] = ' '; 499 conv = 3; 500 501done: 502 /* 503 * The first character cannot be E5, 504 * because that means a deleted entry 505 */ 506 if (dn[0] == 0xe5) 507 dn[0] = SLOT_E5; 508 509 return conv; 510} 511 512/* 513 * Create a Win95 long name directory entry 514 * Note: assumes that the filename is valid, 515 * i.e. doesn't consist solely of blanks and dots 516 */ 517int 518unix2winfn(const u_char *un, size_t unlen, struct winentry *wep, int cnt, 519 int chksum, struct msdosfsmount *pmp) 520{ 521 u_int8_t *wcp; 522 int i, end; 523 u_int16_t code; 524 525 /* 526 * Drop trailing blanks and dots 527 */ 528 unlen = winLenFixup(un, unlen); 529 530 /* 531 * Cut *un for this slot 532 */ 533 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2, 534 pmp->pm_flags, pmp->pm_u2w); 535 536 /* 537 * Initialize winentry to some useful default 538 */ 539 for (wcp = (u_int8_t *)wep, i = sizeof(*wep); --i >= 0; *wcp++ = 0xff); 540 wep->weCnt = cnt; 541 wep->weAttributes = ATTR_WIN95; 542 wep->weReserved1 = 0; 543 wep->weChksum = chksum; 544 wep->weReserved2 = 0; 545 546 /* 547 * Now convert the filename parts 548 */ 549 end = 0; 550 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) { 551 code = unix2winchr(&un, &unlen, 0, pmp); 552 *wcp++ = code; 553 *wcp++ = code >> 8; 554 if (!code) 555 end = WIN_LAST; 556 } 557 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) { 558 code = unix2winchr(&un, &unlen, 0, pmp); 559 *wcp++ = code; 560 *wcp++ = code >> 8; 561 if (!code) 562 end = WIN_LAST; 563 } 564 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) { 565 code = unix2winchr(&un, &unlen, 0, pmp); 566 *wcp++ = code; 567 *wcp++ = code >> 8; 568 if (!code) 569 end = WIN_LAST; 570 } 571 if (*un == '\0') 572 end = WIN_LAST; 573 wep->weCnt |= end; 574 return !end; 575} 576 577/* 578 * Compare our filename to the one in the Win95 entry 579 * Returns the checksum or -1 if no match 580 */ 581int 582winChkName(struct mbnambuf *nbp, const u_char *un, size_t unlen, int chksum, 583 struct msdosfsmount *pmp) 584{ 585 size_t len; 586 u_int16_t c1, c2; 587 u_char *np; 588 struct dirent dirbuf; 589 590 /* 591 * We already have winentry in *nbp. 592 */ 593 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0) 594 return -1; 595 596#ifdef MSDOSFS_DEBUG 597 printf("winChkName(): un=%s:%d,d_name=%s:%d\n", un, unlen, 598 dirbuf.d_name, 599 dirbuf.d_namlen); 600#endif 601 602 /* 603 * Compare the name parts 604 */ 605 len = dirbuf.d_namlen; 606 if (unlen != len) 607 return -2; 608 609 for (np = dirbuf.d_name; unlen > 0 && len > 0;) { 610 /* 611 * Comparison must be case insensitive, because FAT disallows 612 * to look up or create files in case sensitive even when 613 * it's a long file name. 614 */ 615 c1 = unix2winchr((const u_char **)&np, &len, LCASE_BASE, pmp); 616 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp); 617 if (c1 != c2) 618 return -2; 619 } 620 return chksum; 621} 622 623/* 624 * Convert Win95 filename to dirbuf. 625 * Returns the checksum or -1 if impossible 626 */ 627int 628win2unixfn(struct mbnambuf *nbp, struct winentry *wep, int chksum, 629 struct msdosfsmount *pmp) 630{ 631 u_char *c, tmpbuf[5]; 632 u_int8_t *cp; 633 u_int8_t *np, name[WIN_CHARS * 3 + 1]; 634 u_int16_t code; 635 int i; 636 637 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS) 638 || !(wep->weCnt&WIN_CNT)) 639 return -1; 640 641 /* 642 * First compare checksums 643 */ 644 if (wep->weCnt&WIN_LAST) { 645 chksum = wep->weChksum; 646 } else if (chksum != wep->weChksum) 647 chksum = -1; 648 if (chksum == -1) 649 return -1; 650 651 /* 652 * Convert the name parts 653 */ 654 np = name; 655 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) { 656 code = (cp[1] << 8) | cp[0]; 657 switch (code) { 658 case 0: 659 *np = '\0'; 660 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 661 return chksum; 662 case '/': 663 *np = '\0'; 664 return -1; 665 default: 666 c = win2unixchr(tmpbuf, code, pmp); 667 while (*c != '\0') 668 *np++ = *c++; 669 break; 670 } 671 cp += 2; 672 } 673 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) { 674 code = (cp[1] << 8) | cp[0]; 675 switch (code) { 676 case 0: 677 *np = '\0'; 678 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 679 return chksum; 680 case '/': 681 *np = '\0'; 682 return -1; 683 default: 684 c = win2unixchr(tmpbuf, code, pmp); 685 while (*c != '\0') 686 *np++ = *c++; 687 break; 688 } 689 cp += 2; 690 } 691 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) { 692 code = (cp[1] << 8) | cp[0]; 693 switch (code) { 694 case 0: 695 *np = '\0'; 696 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 697 return chksum; 698 case '/': 699 *np = '\0'; 700 return -1; 701 default: 702 c = win2unixchr(tmpbuf, code, pmp); 703 while (*c != '\0') 704 *np++ = *c++; 705 break; 706 } 707 cp += 2; 708 } 709 *np = '\0'; 710 mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1); 711 return chksum; 712} 713 714/* 715 * Compute the unrolled checksum of a DOS filename for Win95 LFN use. 716 */ 717u_int8_t 718winChksum(u_int8_t *name) 719{ 720 int i; 721 u_int8_t s; 722 723 for (s = 0, i = 11; --i >= 0; s += *name++) 724 s = (s << 7)|(s >> 1); 725 return (s); 726} 727 728/* 729 * Determine the number of slots necessary for Win95 names 730 */ 731int 732winSlotCnt(const u_char *un, size_t unlen, struct msdosfsmount *pmp) 733{ 734 size_t wlen; 735 char wn[WIN_MAXLEN * 2 + 1], *wnp; 736 737 unlen = winLenFixup(un, unlen); 738 739 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 740 wlen = WIN_MAXLEN * 2; 741 wnp = wn; 742 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen); 743 if (unlen > 0) 744 return 0; 745 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS); 746 } 747 748 if (unlen > WIN_MAXLEN) 749 return 0; 750 return howmany(unlen, WIN_CHARS); 751} 752 753/* 754 * Determine the number of bytes neccessary for Win95 names 755 */ 756size_t 757winLenFixup(const u_char *un, size_t unlen) 758{ 759 for (un += unlen; unlen > 0; unlen--) 760 if (*--un != ' ' && *un != '.') 761 break; 762 return unlen; 763} 764 765/* 766 * Store an area with multi byte string instr, and returns left 767 * byte of instr and moves pointer forward. The area's size is 768 * inlen or outlen. 769 */ 770static int 771mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle) 772{ 773 char *outp, outstr[outlen * weight + 1]; 774 775 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) { 776 outp = outstr; 777 outlen *= weight; 778 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen); 779 return (inlen); 780 } 781 782 (*instr) += min(inlen, outlen); 783 return (inlen - min(inlen, outlen)); 784} 785 786/* 787 * Convert DOS char to Local char 788 */ 789static u_char * 790dos2unixchr(u_char *outbuf, const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 791{ 792 u_char c, *outp; 793 size_t len, olen; 794 795 outp = outbuf; 796 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 797 olen = len = 4; 798 799 if (lower & (LCASE_BASE | LCASE_EXT)) 800 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr, 801 ilen, (char **)&outp, &olen, KICONV_LOWER); 802 else 803 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr, 804 ilen, (char **)&outp, &olen); 805 len -= olen; 806 807 /* 808 * return '?' if failed to convert 809 */ 810 if (len == 0) { 811 (*ilen)--; 812 (*instr)++; 813 *outp++ = '?'; 814 } 815 } else { 816 (*ilen)--; 817 c = *(*instr)++; 818 c = dos2unix[c]; 819 if (lower & (LCASE_BASE | LCASE_EXT)) 820 c = u2l[c]; 821 *outp++ = c; 822 outbuf[1] = '\0'; 823 } 824 825 *outp = '\0'; 826 outp = outbuf; 827 return (outp); 828} 829 830/* 831 * Convert Local char to DOS char 832 */ 833static u_int16_t 834unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp) 835{ 836 u_char c; 837 char *up, *outp, unicode[3], outbuf[3]; 838 u_int16_t wc; 839 size_t len, ucslen, unixlen, olen; 840 841 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 842 /* 843 * to hide an invisible character, using a unicode filter 844 */ 845 ucslen = 2; 846 len = *ilen; 847 up = unicode; 848 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 849 ilen, &up, &ucslen); 850 unixlen = len - *ilen; 851 852 /* 853 * cannot be converted 854 */ 855 if (unixlen == 0) { 856 (*ilen)--; 857 (*instr)++; 858 return (0); 859 } 860 861 /* 862 * return magic number for ascii char 863 */ 864 if (unixlen == 1) { 865 c = *(*instr -1); 866 if (! (c & 0x80)) { 867 c = unix2dos[c]; 868 if (c <= 2) 869 return (c); 870 } 871 } 872 873 /* 874 * now convert using libiconv 875 */ 876 *instr -= unixlen; 877 *ilen = len; 878 879 olen = len = 2; 880 outp = outbuf; 881 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr, 882 ilen, &outp, &olen, KICONV_FROM_UPPER); 883 len -= olen; 884 885 /* 886 * cannot be converted, but has unicode char, should return magic number 887 */ 888 if (len == 0) { 889 (*ilen) -= unixlen; 890 (*instr) += unixlen; 891 return (1); 892 } 893 894 wc = 0; 895 while(len--) 896 wc |= (*(outp - len - 1) & 0xff) << (len << 3); 897 return (wc); 898 } 899 900 (*ilen)--; 901 c = *(*instr)++; 902 c = l2u[c]; 903 c = unix2dos[c]; 904 return ((u_int16_t)c); 905} 906 907/* 908 * Convert Windows char to Local char 909 */ 910static u_char * 911win2unixchr(u_char *outbuf, u_int16_t wc, struct msdosfsmount *pmp) 912{ 913 u_char *inp, *outp, inbuf[3]; 914 size_t ilen, olen, len; 915 916 outp = outbuf; 917 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 918 inbuf[0] = (u_char)(wc>>8); 919 inbuf[1] = (u_char)wc; 920 inbuf[2] = '\0'; 921 922 ilen = 2; 923 olen = len = 4; 924 inp = inbuf; 925 msdosfs_iconv->convchr(pmp->pm_w2u, (const char **)&inp, &ilen, 926 (char **)&outp, &olen); 927 len -= olen; 928 929 /* 930 * return '?' if failed to convert 931 */ 932 if (len == 0) 933 *outp++ = '?'; 934 } else { 935 *outp++ = (wc & 0xff00) ? '?' : (u_char)(wc & 0xff); 936 } 937 938 *outp = '\0'; 939 outp = outbuf; 940 return (outp); 941} 942 943/* 944 * Convert Local char to Windows char 945 */ 946static u_int16_t 947unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp) 948{ 949 u_char *outp, outbuf[3]; 950 u_int16_t wc; 951 size_t olen; 952 953 if (*ilen == 0) 954 return (0); 955 956 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { 957 outp = outbuf; 958 olen = 2; 959 if (lower & (LCASE_BASE | LCASE_EXT)) 960 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr, 961 ilen, (char **)&outp, &olen, 962 KICONV_FROM_LOWER); 963 else 964 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr, 965 ilen, (char **)&outp, &olen); 966 967 /* 968 * return '0' if end of filename 969 */ 970 if (olen == 2) 971 return (0); 972 973 wc = (outbuf[0]<<8) | outbuf[1]; 974 975 return (wc); 976 } 977 978 (*ilen)--; 979 wc = (*instr)[0]; 980 if (lower & (LCASE_BASE | LCASE_EXT)) 981 wc = u2l[wc]; 982 (*instr)++; 983 return (wc); 984} 985 986/* 987 * Initialize the temporary concatenation buffer. 988 */ 989void 990mbnambuf_init(struct mbnambuf *nbp) 991{ 992 993 nbp->nb_len = 0; 994 nbp->nb_last_id = -1; 995 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0'; 996} 997 998/* 999 * Fill out our concatenation buffer with the given substring, at the offset 1000 * specified by its id. Since this function must be called with ids in 1001 * descending order, we take advantage of the fact that ASCII substrings are 1002 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all 1003 * previous (i.e. higher id) substrings upwards to make room for this one. 1004 * This only penalizes portions of substrings that contain more than 1005 * WIN_CHARS bytes when they are first encountered. 1006 */ 1007void 1008mbnambuf_write(struct mbnambuf *nbp, char *name, int id) 1009{ 1010 char *slot; 1011 size_t count, newlen; 1012 1013 if (nbp->nb_len != 0 && id != nbp->nb_last_id - 1) { 1014#ifdef MSDOSFS_DEBUG 1015 printf("msdosfs: non-decreasing id: id %d, last id %d\n", 1016 id, nbp->nb_last_id); 1017#endif 1018 return; 1019 } 1020 1021 /* Will store this substring in a WIN_CHARS-aligned slot. */ 1022 slot = &nbp->nb_buf[id * WIN_CHARS]; 1023 count = strlen(name); 1024 newlen = nbp->nb_len + count; 1025 if (newlen > WIN_MAXLEN || newlen > MAXNAMLEN) { 1026#ifdef MSDOSFS_DEBUG 1027 printf("msdosfs: file name length %zu too large\n", newlen); 1028#endif 1029 return; 1030 } 1031 1032 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */ 1033 if (count > WIN_CHARS && nbp->nb_len != 0) 1034 bcopy(slot + WIN_CHARS, slot + count, nbp->nb_len); 1035 1036 /* Copy in the substring to its slot and update length so far. */ 1037 bcopy(name, slot, count); 1038 nbp->nb_len = newlen; 1039 nbp->nb_last_id = id; 1040} 1041 1042/* 1043 * Take the completed string and use it to setup the struct dirent. 1044 * Be sure to always nul-terminate the d_name and then copy the string 1045 * from our buffer. Note that this function assumes the full string has 1046 * been reassembled in the buffer. If it's called before all substrings 1047 * have been written via mbnambuf_write(), the result will be incorrect. 1048 */ 1049char * 1050mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp) 1051{ 1052 1053 if (nbp->nb_len > sizeof(dp->d_name) - 1) { 1054 mbnambuf_init(nbp); 1055 return (NULL); 1056 } 1057 bcopy(&nbp->nb_buf[0], dp->d_name, nbp->nb_len); 1058 dp->d_name[nbp->nb_len] = '\0'; 1059 dp->d_namlen = nbp->nb_len; 1060 1061 mbnambuf_init(nbp); 1062 return (dp->d_name); 1063} 1064