base64.c revision 126274
1/* OPENBSD ORIGINAL: lib/libc/net/base64.c */ 2 3/* $OpenBSD: base64.c,v 1.4 2002/01/02 23:00:10 deraadt Exp $ */ 4 5/* 6 * Copyright (c) 1996 by Internet Software Consortium. 7 * 8 * Permission to use, copy, modify, and distribute this software for any 9 * purpose with or without fee is hereby granted, provided that the above 10 * copyright notice and this permission notice appear in all copies. 11 * 12 * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS 13 * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES 14 * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE 15 * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL 16 * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR 17 * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS 18 * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS 19 * SOFTWARE. 20 */ 21 22/* 23 * Portions Copyright (c) 1995 by International Business Machines, Inc. 24 * 25 * International Business Machines, Inc. (hereinafter called IBM) grants 26 * permission under its copyrights to use, copy, modify, and distribute this 27 * Software with or without fee, provided that the above copyright notice and 28 * all paragraphs of this notice appear in all copies, and that the name of IBM 29 * not be used in connection with the marketing of any product incorporating 30 * the Software or modifications thereof, without specific, written prior 31 * permission. 32 * 33 * To the extent it has a right to do so, IBM grants an immunity from suit 34 * under its patents, if any, for the use, sale or manufacture of products to 35 * the extent that such products are used for performing Domain Name System 36 * dynamic updates in TCP/IP networks by means of the Software. No immunity is 37 * granted for any product per se or for any other function of any product. 38 * 39 * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES, 40 * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A 41 * PARTICULAR PURPOSE. IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL, 42 * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING 43 * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN 44 * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES. 45 */ 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 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 int tarindex, state, ch; 210 char *pos; 211 212 state = 0; 213 tarindex = 0; 214 215 while ((ch = *src++) != '\0') { 216 if (isspace(ch)) /* Skip whitespace anywhere. */ 217 continue; 218 219 if (ch == Pad64) 220 break; 221 222 pos = strchr(Base64, ch); 223 if (pos == 0) /* A non-base64 character. */ 224 return (-1); 225 226 switch (state) { 227 case 0: 228 if (target) { 229 if (tarindex >= targsize) 230 return (-1); 231 target[tarindex] = (pos - Base64) << 2; 232 } 233 state = 1; 234 break; 235 case 1: 236 if (target) { 237 if (tarindex + 1 >= targsize) 238 return (-1); 239 target[tarindex] |= (pos - Base64) >> 4; 240 target[tarindex+1] = ((pos - Base64) & 0x0f) 241 << 4 ; 242 } 243 tarindex++; 244 state = 2; 245 break; 246 case 2: 247 if (target) { 248 if (tarindex + 1 >= targsize) 249 return (-1); 250 target[tarindex] |= (pos - Base64) >> 2; 251 target[tarindex+1] = ((pos - Base64) & 0x03) 252 << 6; 253 } 254 tarindex++; 255 state = 3; 256 break; 257 case 3: 258 if (target) { 259 if (tarindex >= targsize) 260 return (-1); 261 target[tarindex] |= (pos - Base64); 262 } 263 tarindex++; 264 state = 0; 265 break; 266 } 267 } 268 269 /* 270 * We are done decoding Base-64 chars. Let's see if we ended 271 * on a byte boundary, and/or with erroneous trailing characters. 272 */ 273 274 if (ch == Pad64) { /* We got a pad char. */ 275 ch = *src++; /* Skip it, get next. */ 276 switch (state) { 277 case 0: /* Invalid = in first position */ 278 case 1: /* Invalid = in second position */ 279 return (-1); 280 281 case 2: /* Valid, means one byte of info */ 282 /* Skip any number of spaces. */ 283 for (; ch != '\0'; ch = *src++) 284 if (!isspace(ch)) 285 break; 286 /* Make sure there is another trailing = sign. */ 287 if (ch != Pad64) 288 return (-1); 289 ch = *src++; /* Skip the = */ 290 /* Fall through to "single trailing =" case. */ 291 /* FALLTHROUGH */ 292 293 case 3: /* Valid, means two bytes of info */ 294 /* 295 * We know this char is an =. Is there anything but 296 * whitespace after it? 297 */ 298 for (; ch != '\0'; ch = *src++) 299 if (!isspace(ch)) 300 return (-1); 301 302 /* 303 * Now make sure for cases 2 and 3 that the "extra" 304 * bits that slopped past the last full byte were 305 * zeros. If we don't check them, they become a 306 * subliminal channel. 307 */ 308 if (target && target[tarindex] != 0) 309 return (-1); 310 } 311 } else { 312 /* 313 * We ended by seeing the end of the string. Make sure we 314 * have no partial bytes lying around. 315 */ 316 if (state != 0) 317 return (-1); 318 } 319 320 return (tarindex); 321} 322 323#endif /* !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON) */ 324#endif 325