base64.c revision 157017
1/* $OpenBSD: base64.c,v 1.4 2002/01/02 23:00:10 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/* OPENBSD ORIGINAL: lib/libc/net/base64.c */ 46 47#include "includes.h" 48 49#if (!defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)) || (!defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON)) 50 51#include <sys/types.h> 52#include <sys/param.h> 53#include <sys/socket.h> 54#include <netinet/in.h> 55#include <arpa/inet.h> 56 57#include <ctype.h> 58#include <stdio.h> 59 60#include <stdlib.h> 61#include <string.h> 62 63#include "base64.h" 64 65/* XXX abort illegal in library */ 66#define Assert(Cond) if (!(Cond)) abort() 67 68static const char Base64[] = 69 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; 70static const char Pad64 = '='; 71 72/* (From RFC1521 and draft-ietf-dnssec-secext-03.txt) 73 The following encoding technique is taken from RFC 1521 by Borenstein 74 and Freed. It is reproduced here in a slightly edited form for 75 convenience. 76 77 A 65-character subset of US-ASCII is used, enabling 6 bits to be 78 represented per printable character. (The extra 65th character, "=", 79 is used to signify a special processing function.) 80 81 The encoding process represents 24-bit groups of input bits as output 82 strings of 4 encoded characters. Proceeding from left to right, a 83 24-bit input group is formed by concatenating 3 8-bit input groups. 84 These 24 bits are then treated as 4 concatenated 6-bit groups, each 85 of which is translated into a single digit in the base64 alphabet. 86 87 Each 6-bit group is used as an index into an array of 64 printable 88 characters. The character referenced by the index is placed in the 89 output string. 90 91 Table 1: The Base64 Alphabet 92 93 Value Encoding Value Encoding Value Encoding Value Encoding 94 0 A 17 R 34 i 51 z 95 1 B 18 S 35 j 52 0 96 2 C 19 T 36 k 53 1 97 3 D 20 U 37 l 54 2 98 4 E 21 V 38 m 55 3 99 5 F 22 W 39 n 56 4 100 6 G 23 X 40 o 57 5 101 7 H 24 Y 41 p 58 6 102 8 I 25 Z 42 q 59 7 103 9 J 26 a 43 r 60 8 104 10 K 27 b 44 s 61 9 105 11 L 28 c 45 t 62 + 106 12 M 29 d 46 u 63 / 107 13 N 30 e 47 v 108 14 O 31 f 48 w (pad) = 109 15 P 32 g 49 x 110 16 Q 33 h 50 y 111 112 Special processing is performed if fewer than 24 bits are available 113 at the end of the data being encoded. A full encoding quantum is 114 always completed at the end of a quantity. When fewer than 24 input 115 bits are available in an input group, zero bits are added (on the 116 right) to form an integral number of 6-bit groups. Padding at the 117 end of the data is performed using the '=' character. 118 119 Since all base64 input is an integral number of octets, only the 120 ------------------------------------------------- 121 following cases can arise: 122 123 (1) the final quantum of encoding input is an integral 124 multiple of 24 bits; here, the final unit of encoded 125 output will be an integral multiple of 4 characters 126 with no "=" padding, 127 (2) the final quantum of encoding input is exactly 8 bits; 128 here, the final unit of encoded output will be two 129 characters followed by two "=" padding characters, or 130 (3) the final quantum of encoding input is exactly 16 bits; 131 here, the final unit of encoded output will be three 132 characters followed by one "=" padding character. 133 */ 134 135#if !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) 136int 137b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) 138{ 139 size_t datalength = 0; 140 u_char input[3]; 141 u_char output[4]; 142 u_int i; 143 144 while (2 < srclength) { 145 input[0] = *src++; 146 input[1] = *src++; 147 input[2] = *src++; 148 srclength -= 3; 149 150 output[0] = input[0] >> 2; 151 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4); 152 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6); 153 output[3] = input[2] & 0x3f; 154 Assert(output[0] < 64); 155 Assert(output[1] < 64); 156 Assert(output[2] < 64); 157 Assert(output[3] < 64); 158 159 if (datalength + 4 > targsize) 160 return (-1); 161 target[datalength++] = Base64[output[0]]; 162 target[datalength++] = Base64[output[1]]; 163 target[datalength++] = Base64[output[2]]; 164 target[datalength++] = Base64[output[3]]; 165 } 166 167 /* Now we worry about padding. */ 168 if (0 != srclength) { 169 /* Get what's left. */ 170 input[0] = input[1] = input[2] = '\0'; 171 for (i = 0; i < srclength; i++) 172 input[i] = *src++; 173 174 output[0] = input[0] >> 2; 175 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4); 176 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6); 177 Assert(output[0] < 64); 178 Assert(output[1] < 64); 179 Assert(output[2] < 64); 180 181 if (datalength + 4 > targsize) 182 return (-1); 183 target[datalength++] = Base64[output[0]]; 184 target[datalength++] = Base64[output[1]]; 185 if (srclength == 1) 186 target[datalength++] = Pad64; 187 else 188 target[datalength++] = Base64[output[2]]; 189 target[datalength++] = Pad64; 190 } 191 if (datalength >= targsize) 192 return (-1); 193 target[datalength] = '\0'; /* Returned value doesn't count \0. */ 194 return (datalength); 195} 196#endif /* !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) */ 197 198#if !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON) 199 200/* skips all whitespace anywhere. 201 converts characters, four at a time, starting at (or after) 202 src from base - 64 numbers into three 8 bit bytes in the target area. 203 it returns the number of data bytes stored at the target, or -1 on error. 204 */ 205 206int 207b64_pton(char const *src, u_char *target, size_t targsize) 208{ 209 u_int tarindex, state; 210 int ch; 211 char *pos; 212 213 state = 0; 214 tarindex = 0; 215 216 while ((ch = *src++) != '\0') { 217 if (isspace(ch)) /* Skip whitespace anywhere. */ 218 continue; 219 220 if (ch == Pad64) 221 break; 222 223 pos = strchr(Base64, ch); 224 if (pos == 0) /* A non-base64 character. */ 225 return (-1); 226 227 switch (state) { 228 case 0: 229 if (target) { 230 if (tarindex >= targsize) 231 return (-1); 232 target[tarindex] = (pos - Base64) << 2; 233 } 234 state = 1; 235 break; 236 case 1: 237 if (target) { 238 if (tarindex + 1 >= targsize) 239 return (-1); 240 target[tarindex] |= (pos - Base64) >> 4; 241 target[tarindex+1] = ((pos - Base64) & 0x0f) 242 << 4 ; 243 } 244 tarindex++; 245 state = 2; 246 break; 247 case 2: 248 if (target) { 249 if (tarindex + 1 >= targsize) 250 return (-1); 251 target[tarindex] |= (pos - Base64) >> 2; 252 target[tarindex+1] = ((pos - Base64) & 0x03) 253 << 6; 254 } 255 tarindex++; 256 state = 3; 257 break; 258 case 3: 259 if (target) { 260 if (tarindex >= targsize) 261 return (-1); 262 target[tarindex] |= (pos - Base64); 263 } 264 tarindex++; 265 state = 0; 266 break; 267 } 268 } 269 270 /* 271 * We are done decoding Base-64 chars. Let's see if we ended 272 * on a byte boundary, and/or with erroneous trailing characters. 273 */ 274 275 if (ch == Pad64) { /* We got a pad char. */ 276 ch = *src++; /* Skip it, get next. */ 277 switch (state) { 278 case 0: /* Invalid = in first position */ 279 case 1: /* Invalid = in second position */ 280 return (-1); 281 282 case 2: /* Valid, means one byte of info */ 283 /* Skip any number of spaces. */ 284 for (; ch != '\0'; ch = *src++) 285 if (!isspace(ch)) 286 break; 287 /* Make sure there is another trailing = sign. */ 288 if (ch != Pad64) 289 return (-1); 290 ch = *src++; /* Skip the = */ 291 /* Fall through to "single trailing =" case. */ 292 /* FALLTHROUGH */ 293 294 case 3: /* Valid, means two bytes of info */ 295 /* 296 * We know this char is an =. Is there anything but 297 * whitespace after it? 298 */ 299 for (; ch != '\0'; ch = *src++) 300 if (!isspace(ch)) 301 return (-1); 302 303 /* 304 * Now make sure for cases 2 and 3 that the "extra" 305 * bits that slopped past the last full byte were 306 * zeros. If we don't check them, they become a 307 * subliminal channel. 308 */ 309 if (target && target[tarindex] != 0) 310 return (-1); 311 } 312 } else { 313 /* 314 * We ended by seeing the end of the string. Make sure we 315 * have no partial bytes lying around. 316 */ 317 if (state != 0) 318 return (-1); 319 } 320 321 return (tarindex); 322} 323 324#endif /* !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON) */ 325#endif 326