Damien Miller | e247cc4 | 2000-05-07 12:03:14 +1000 | [diff] [blame^] | 1 | /* $OpenBSD: base64.c,v 1.3 1997/11/08 20:46:55 deraadt Exp $ */ |
| 2 | |
| 3 | /* |
| 4 | * Copyright (c) 1996 by Internet Software Consortium. |
| 5 | * |
| 6 | * Permission to use, copy, modify, and distribute this software for any |
| 7 | * purpose with or without fee is hereby granted, provided that the above |
| 8 | * copyright notice and this permission notice appear in all copies. |
| 9 | * |
| 10 | * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS |
| 11 | * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES |
| 12 | * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE |
| 13 | * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL |
| 14 | * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR |
| 15 | * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS |
| 16 | * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS |
| 17 | * SOFTWARE. |
| 18 | */ |
| 19 | |
| 20 | /* |
| 21 | * Portions Copyright (c) 1995 by International Business Machines, Inc. |
| 22 | * |
| 23 | * International Business Machines, Inc. (hereinafter called IBM) grants |
| 24 | * permission under its copyrights to use, copy, modify, and distribute this |
| 25 | * Software with or without fee, provided that the above copyright notice and |
| 26 | * all paragraphs of this notice appear in all copies, and that the name of IBM |
| 27 | * not be used in connection with the marketing of any product incorporating |
| 28 | * the Software or modifications thereof, without specific, written prior |
| 29 | * permission. |
| 30 | * |
| 31 | * To the extent it has a right to do so, IBM grants an immunity from suit |
| 32 | * under its patents, if any, for the use, sale or manufacture of products to |
| 33 | * the extent that such products are used for performing Domain Name System |
| 34 | * dynamic updates in TCP/IP networks by means of the Software. No immunity is |
| 35 | * granted for any product per se or for any other function of any product. |
| 36 | * |
| 37 | * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES, |
| 38 | * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A |
| 39 | * PARTICULAR PURPOSE. IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL, |
| 40 | * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING |
| 41 | * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN |
| 42 | * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES. |
| 43 | */ |
| 44 | |
| 45 | #include "config.h" |
| 46 | |
| 47 | #ifndef HAVE_B64_NTOP |
| 48 | |
| 49 | #include <sys/types.h> |
| 50 | #include <sys/param.h> |
| 51 | #include <sys/socket.h> |
| 52 | #include <netinet/in.h> |
| 53 | #include <arpa/inet.h> |
| 54 | |
| 55 | #include <ctype.h> |
| 56 | #include <resolv.h> |
| 57 | #include <stdio.h> |
| 58 | |
| 59 | #include <stdlib.h> |
| 60 | #include <string.h> |
| 61 | |
| 62 | #define Assert(Cond) if (!(Cond)) abort() |
| 63 | |
| 64 | static const char Base64[] = |
| 65 | "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; |
| 66 | static const char Pad64 = '='; |
| 67 | |
| 68 | /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt) |
| 69 | The following encoding technique is taken from RFC 1521 by Borenstein |
| 70 | and Freed. It is reproduced here in a slightly edited form for |
| 71 | convenience. |
| 72 | |
| 73 | A 65-character subset of US-ASCII is used, enabling 6 bits to be |
| 74 | represented per printable character. (The extra 65th character, "=", |
| 75 | is used to signify a special processing function.) |
| 76 | |
| 77 | The encoding process represents 24-bit groups of input bits as output |
| 78 | strings of 4 encoded characters. Proceeding from left to right, a |
| 79 | 24-bit input group is formed by concatenating 3 8-bit input groups. |
| 80 | These 24 bits are then treated as 4 concatenated 6-bit groups, each |
| 81 | of which is translated into a single digit in the base64 alphabet. |
| 82 | |
| 83 | Each 6-bit group is used as an index into an array of 64 printable |
| 84 | characters. The character referenced by the index is placed in the |
| 85 | output string. |
| 86 | |
| 87 | Table 1: The Base64 Alphabet |
| 88 | |
| 89 | Value Encoding Value Encoding Value Encoding Value Encoding |
| 90 | 0 A 17 R 34 i 51 z |
| 91 | 1 B 18 S 35 j 52 0 |
| 92 | 2 C 19 T 36 k 53 1 |
| 93 | 3 D 20 U 37 l 54 2 |
| 94 | 4 E 21 V 38 m 55 3 |
| 95 | 5 F 22 W 39 n 56 4 |
| 96 | 6 G 23 X 40 o 57 5 |
| 97 | 7 H 24 Y 41 p 58 6 |
| 98 | 8 I 25 Z 42 q 59 7 |
| 99 | 9 J 26 a 43 r 60 8 |
| 100 | 10 K 27 b 44 s 61 9 |
| 101 | 11 L 28 c 45 t 62 + |
| 102 | 12 M 29 d 46 u 63 / |
| 103 | 13 N 30 e 47 v |
| 104 | 14 O 31 f 48 w (pad) = |
| 105 | 15 P 32 g 49 x |
| 106 | 16 Q 33 h 50 y |
| 107 | |
| 108 | Special processing is performed if fewer than 24 bits are available |
| 109 | at the end of the data being encoded. A full encoding quantum is |
| 110 | always completed at the end of a quantity. When fewer than 24 input |
| 111 | bits are available in an input group, zero bits are added (on the |
| 112 | right) to form an integral number of 6-bit groups. Padding at the |
| 113 | end of the data is performed using the '=' character. |
| 114 | |
| 115 | Since all base64 input is an integral number of octets, only the |
| 116 | ------------------------------------------------- |
| 117 | following cases can arise: |
| 118 | |
| 119 | (1) the final quantum of encoding input is an integral |
| 120 | multiple of 24 bits; here, the final unit of encoded |
| 121 | output will be an integral multiple of 4 characters |
| 122 | with no "=" padding, |
| 123 | (2) the final quantum of encoding input is exactly 8 bits; |
| 124 | here, the final unit of encoded output will be two |
| 125 | characters followed by two "=" padding characters, or |
| 126 | (3) the final quantum of encoding input is exactly 16 bits; |
| 127 | here, the final unit of encoded output will be three |
| 128 | characters followed by one "=" padding character. |
| 129 | */ |
| 130 | |
| 131 | int |
| 132 | b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) |
| 133 | { |
| 134 | size_t datalength = 0; |
| 135 | u_char input[3]; |
| 136 | u_char output[4]; |
| 137 | int i; |
| 138 | |
| 139 | while (2 < srclength) { |
| 140 | input[0] = *src++; |
| 141 | input[1] = *src++; |
| 142 | input[2] = *src++; |
| 143 | srclength -= 3; |
| 144 | |
| 145 | output[0] = input[0] >> 2; |
| 146 | output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4); |
| 147 | output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6); |
| 148 | output[3] = input[2] & 0x3f; |
| 149 | Assert(output[0] < 64); |
| 150 | Assert(output[1] < 64); |
| 151 | Assert(output[2] < 64); |
| 152 | Assert(output[3] < 64); |
| 153 | |
| 154 | if (datalength + 4 > targsize) |
| 155 | return (-1); |
| 156 | target[datalength++] = Base64[output[0]]; |
| 157 | target[datalength++] = Base64[output[1]]; |
| 158 | target[datalength++] = Base64[output[2]]; |
| 159 | target[datalength++] = Base64[output[3]]; |
| 160 | } |
| 161 | |
| 162 | /* Now we worry about padding. */ |
| 163 | if (0 != srclength) { |
| 164 | /* Get what's left. */ |
| 165 | input[0] = input[1] = input[2] = '\0'; |
| 166 | for (i = 0; i < srclength; i++) |
| 167 | input[i] = *src++; |
| 168 | |
| 169 | output[0] = input[0] >> 2; |
| 170 | output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4); |
| 171 | output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6); |
| 172 | Assert(output[0] < 64); |
| 173 | Assert(output[1] < 64); |
| 174 | Assert(output[2] < 64); |
| 175 | |
| 176 | if (datalength + 4 > targsize) |
| 177 | return (-1); |
| 178 | target[datalength++] = Base64[output[0]]; |
| 179 | target[datalength++] = Base64[output[1]]; |
| 180 | if (srclength == 1) |
| 181 | target[datalength++] = Pad64; |
| 182 | else |
| 183 | target[datalength++] = Base64[output[2]]; |
| 184 | target[datalength++] = Pad64; |
| 185 | } |
| 186 | if (datalength >= targsize) |
| 187 | return (-1); |
| 188 | target[datalength] = '\0'; /* Returned value doesn't count \0. */ |
| 189 | return (datalength); |
| 190 | } |
| 191 | |
| 192 | /* skips all whitespace anywhere. |
| 193 | converts characters, four at a time, starting at (or after) |
| 194 | src from base - 64 numbers into three 8 bit bytes in the target area. |
| 195 | it returns the number of data bytes stored at the target, or -1 on error. |
| 196 | */ |
| 197 | |
| 198 | int |
| 199 | b64_pton(char const *src, u_char *target, size_t targsize) |
| 200 | { |
| 201 | int tarindex, state, ch; |
| 202 | char *pos; |
| 203 | |
| 204 | state = 0; |
| 205 | tarindex = 0; |
| 206 | |
| 207 | while ((ch = *src++) != '\0') { |
| 208 | if (isspace(ch)) /* Skip whitespace anywhere. */ |
| 209 | continue; |
| 210 | |
| 211 | if (ch == Pad64) |
| 212 | break; |
| 213 | |
| 214 | pos = strchr(Base64, ch); |
| 215 | if (pos == 0) /* A non-base64 character. */ |
| 216 | return (-1); |
| 217 | |
| 218 | switch (state) { |
| 219 | case 0: |
| 220 | if (target) { |
| 221 | if (tarindex >= targsize) |
| 222 | return (-1); |
| 223 | target[tarindex] = (pos - Base64) << 2; |
| 224 | } |
| 225 | state = 1; |
| 226 | break; |
| 227 | case 1: |
| 228 | if (target) { |
| 229 | if (tarindex + 1 >= targsize) |
| 230 | return (-1); |
| 231 | target[tarindex] |= (pos - Base64) >> 4; |
| 232 | target[tarindex+1] = ((pos - Base64) & 0x0f) |
| 233 | << 4 ; |
| 234 | } |
| 235 | tarindex++; |
| 236 | state = 2; |
| 237 | break; |
| 238 | case 2: |
| 239 | if (target) { |
| 240 | if (tarindex + 1 >= targsize) |
| 241 | return (-1); |
| 242 | target[tarindex] |= (pos - Base64) >> 2; |
| 243 | target[tarindex+1] = ((pos - Base64) & 0x03) |
| 244 | << 6; |
| 245 | } |
| 246 | tarindex++; |
| 247 | state = 3; |
| 248 | break; |
| 249 | case 3: |
| 250 | if (target) { |
| 251 | if (tarindex >= targsize) |
| 252 | return (-1); |
| 253 | target[tarindex] |= (pos - Base64); |
| 254 | } |
| 255 | tarindex++; |
| 256 | state = 0; |
| 257 | break; |
| 258 | } |
| 259 | } |
| 260 | |
| 261 | /* |
| 262 | * We are done decoding Base-64 chars. Let's see if we ended |
| 263 | * on a byte boundary, and/or with erroneous trailing characters. |
| 264 | */ |
| 265 | |
| 266 | if (ch == Pad64) { /* We got a pad char. */ |
| 267 | ch = *src++; /* Skip it, get next. */ |
| 268 | switch (state) { |
| 269 | case 0: /* Invalid = in first position */ |
| 270 | case 1: /* Invalid = in second position */ |
| 271 | return (-1); |
| 272 | |
| 273 | case 2: /* Valid, means one byte of info */ |
| 274 | /* Skip any number of spaces. */ |
| 275 | for (; ch != '\0'; ch = *src++) |
| 276 | if (!isspace(ch)) |
| 277 | break; |
| 278 | /* Make sure there is another trailing = sign. */ |
| 279 | if (ch != Pad64) |
| 280 | return (-1); |
| 281 | ch = *src++; /* Skip the = */ |
| 282 | /* Fall through to "single trailing =" case. */ |
| 283 | /* FALLTHROUGH */ |
| 284 | |
| 285 | case 3: /* Valid, means two bytes of info */ |
| 286 | /* |
| 287 | * We know this char is an =. Is there anything but |
| 288 | * whitespace after it? |
| 289 | */ |
| 290 | for (; ch != '\0'; ch = *src++) |
| 291 | if (!isspace(ch)) |
| 292 | return (-1); |
| 293 | |
| 294 | /* |
| 295 | * Now make sure for cases 2 and 3 that the "extra" |
| 296 | * bits that slopped past the last full byte were |
| 297 | * zeros. If we don't check them, they become a |
| 298 | * subliminal channel. |
| 299 | */ |
| 300 | if (target && target[tarindex] != 0) |
| 301 | return (-1); |
| 302 | } |
| 303 | } else { |
| 304 | /* |
| 305 | * We ended by seeing the end of the string. Make sure we |
| 306 | * have no partial bytes lying around. |
| 307 | */ |
| 308 | if (state != 0) |
| 309 | return (-1); |
| 310 | } |
| 311 | |
| 312 | return (tarindex); |
| 313 | } |
| 314 | |
| 315 | #endif /* HAVE_B64_NTOP */ |