1 /* $NetBSD: base64.c,v 1.16 2014/11/24 15:43:21 christos Exp $ */ 2 3 /* 4 * Copyright (c) 2004 by Internet Systems Consortium, Inc. ("ISC") 5 * Copyright (c) 1996-1999 by Internet Software Consortium. 6 * 7 * Permission to use, copy, modify, and distribute this software for any 8 * purpose with or without fee is hereby granted, provided that the above 9 * copyright notice and this permission notice appear in all copies. 10 * 11 * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES 12 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF 13 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR 14 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 15 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN 16 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT 17 * OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS 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 <sys/cdefs.h> 46 #if defined(LIBC_SCCS) && !defined(lint) 47 #if 0 48 static const char rcsid[] = "Id: base64.c,v 1.4 2005/04/27 04:56:34 sra Exp"; 49 #else 50 __RCSID("$NetBSD: base64.c,v 1.16 2014/11/24 15:43:21 christos Exp $"); 51 #endif 52 #endif /* LIBC_SCCS and not lint */ 53 #include "namespace.h" 54 55 #include "port_before.h" 56 57 #include <sys/types.h> 58 #include <sys/param.h> 59 #include <sys/socket.h> 60 61 #include <netinet/in.h> 62 #include <arpa/inet.h> 63 #include <arpa/nameser.h> 64 65 #include <assert.h> 66 #include <ctype.h> 67 #include <resolv.h> 68 #include <stdio.h> 69 #include <stdlib.h> 70 #include <string.h> 71 72 #include "port_after.h" 73 74 #define Assert(Cond) if (!(Cond)) abort() 75 76 static const char Base64[] = 77 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; 78 static const char Pad64 = '='; 79 80 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt) 81 The following encoding technique is taken from RFC1521 by Borenstein 82 and Freed. It is reproduced here in a slightly edited form for 83 convenience. 84 85 A 65-character subset of US-ASCII is used, enabling 6 bits to be 86 represented per printable character. (The extra 65th character, "=", 87 is used to signify a special processing function.) 88 89 The encoding process represents 24-bit groups of input bits as output 90 strings of 4 encoded characters. Proceeding from left to right, a 91 24-bit input group is formed by concatenating 3 8-bit input groups. 92 These 24 bits are then treated as 4 concatenated 6-bit groups, each 93 of which is translated into a single digit in the base64 alphabet. 94 95 Each 6-bit group is used as an index into an array of 64 printable 96 characters. The character referenced by the index is placed in the 97 output string. 98 99 Table 1: The Base64 Alphabet 100 101 Value Encoding Value Encoding Value Encoding Value Encoding 102 0 A 17 R 34 i 51 z 103 1 B 18 S 35 j 52 0 104 2 C 19 T 36 k 53 1 105 3 D 20 U 37 l 54 2 106 4 E 21 V 38 m 55 3 107 5 F 22 W 39 n 56 4 108 6 G 23 X 40 o 57 5 109 7 H 24 Y 41 p 58 6 110 8 I 25 Z 42 q 59 7 111 9 J 26 a 43 r 60 8 112 10 K 27 b 44 s 61 9 113 11 L 28 c 45 t 62 + 114 12 M 29 d 46 u 63 / 115 13 N 30 e 47 v 116 14 O 31 f 48 w (pad) = 117 15 P 32 g 49 x 118 16 Q 33 h 50 y 119 120 Special processing is performed if fewer than 24 bits are available 121 at the end of the data being encoded. A full encoding quantum is 122 always completed at the end of a quantity. When fewer than 24 input 123 bits are available in an input group, zero bits are added (on the 124 right) to form an integral number of 6-bit groups. Padding at the 125 end of the data is performed using the '=' character. 126 127 Since all base64 input is an integral number of octets, only the 128 ------------------------------------------------- 129 following cases can arise: 130 131 (1) the final quantum of encoding input is an integral 132 multiple of 24 bits; here, the final unit of encoded 133 output will be an integral multiple of 4 characters 134 with no "=" padding, 135 (2) the final quantum of encoding input is exactly 8 bits; 136 here, the final unit of encoded output will be two 137 characters followed by two "=" padding characters, or 138 (3) the final quantum of encoding input is exactly 16 bits; 139 here, the final unit of encoded output will be three 140 characters followed by one "=" padding character. 141 */ 142 143 int 144 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) { 145 size_t datalength = 0; 146 u_char input[3]; 147 u_char output[4]; 148 size_t i; 149 150 _DIAGASSERT(src != NULL); 151 _DIAGASSERT(target != NULL); 152 153 while (2U < srclength) { 154 input[0] = *src++; 155 input[1] = *src++; 156 input[2] = *src++; 157 srclength -= 3; 158 159 output[0] = (uint32_t)input[0] >> 2; 160 output[1] = ((uint32_t)(input[0] & 0x03) << 4) + 161 ((uint32_t)input[1] >> 4); 162 output[2] = ((uint32_t)(input[1] & 0x0f) << 2) + 163 ((uint32_t)input[2] >> 6); 164 output[3] = input[2] & 0x3f; 165 Assert(output[0] < 64); 166 Assert(output[1] < 64); 167 Assert(output[2] < 64); 168 Assert(output[3] < 64); 169 170 if (datalength + 4 > targsize) 171 return -1; 172 target[datalength++] = Base64[output[0]]; 173 target[datalength++] = Base64[output[1]]; 174 target[datalength++] = Base64[output[2]]; 175 target[datalength++] = Base64[output[3]]; 176 } 177 178 /* Now we worry about padding. */ 179 if (0U != srclength) { 180 /* Get what's left. */ 181 input[0] = input[1] = input[2] = '\0'; 182 for (i = 0; i < srclength; i++) 183 input[i] = *src++; 184 185 output[0] = (uint32_t)input[0] >> 2; 186 output[1] = ((uint32_t)(input[0] & 0x03) << 4) + 187 ((uint32_t)input[1] >> 4); 188 output[2] = ((uint32_t)(input[1] & 0x0f) << 2) + 189 ((uint32_t)input[2] >> 6); 190 Assert(output[0] < 64); 191 Assert(output[1] < 64); 192 Assert(output[2] < 64); 193 194 if (datalength + 4 > targsize) 195 return -1; 196 target[datalength++] = Base64[output[0]]; 197 target[datalength++] = Base64[output[1]]; 198 if (srclength == 1U) 199 target[datalength++] = Pad64; 200 else 201 target[datalength++] = Base64[output[2]]; 202 target[datalength++] = Pad64; 203 } 204 if (datalength >= targsize) 205 return -1; 206 target[datalength] = '\0'; /*%< Returned value doesn't count \\0. */ 207 assert (datalength <= (size_t)INT_MAX); 208 return (int)datalength; 209 } 210 211 /* skips all whitespace anywhere. 212 converts characters, four at a time, starting at (or after) 213 src from base - 64 numbers into three 8 bit bytes in the target area. 214 it returns the number of data bytes stored at the target, or -1 on error. 215 */ 216 217 int 218 b64_pton(char const *src, u_char *target, size_t targsize) 219 { 220 size_t tarindex; 221 int state, ch; 222 u_char nextbyte; 223 char *pos; 224 225 _DIAGASSERT(src != NULL); 226 _DIAGASSERT(target != NULL); 227 228 state = 0; 229 tarindex = 0; 230 231 while ((ch = (u_char) *src++) != '\0') { 232 if (isspace(ch)) /*%< Skip whitespace anywhere. */ 233 continue; 234 235 if (ch == Pad64) 236 break; 237 238 pos = strchr(Base64, ch); 239 if (pos == NULL) /*%< A non-base64 character. */ 240 return -1; 241 242 switch (state) { 243 case 0: 244 if (target) { 245 if (tarindex >= targsize) 246 return -1; 247 target[tarindex] = (u_char)(pos - Base64) << 2; 248 } 249 state = 1; 250 break; 251 case 1: 252 if (target) { 253 if (tarindex >= targsize) 254 return -1; 255 target[tarindex] |= 256 (uint32_t)(pos - Base64) >> 4; 257 nextbyte = (u_char)((pos - Base64) & 0x0f) << 4; 258 if (tarindex + 1 < targsize) 259 target[tarindex + 1] = nextbyte; 260 else if (nextbyte) 261 return -1; 262 } 263 tarindex++; 264 state = 2; 265 break; 266 case 2: 267 if (target) { 268 if (tarindex >= targsize) 269 return -1; 270 target[tarindex] |= 271 (uint32_t)(pos - Base64) >> 2; 272 nextbyte = (u_char)((pos - Base64) & 0x03) << 6; 273 if (tarindex + 1 < targsize) 274 target[tarindex + 1] = nextbyte; 275 else if (nextbyte) 276 return -1; 277 } 278 tarindex++; 279 state = 3; 280 break; 281 case 3: 282 if (target) { 283 if ((size_t)tarindex >= targsize) 284 return -1; 285 target[tarindex] |= (u_char)(pos - Base64); 286 } 287 tarindex++; 288 state = 0; 289 break; 290 default: 291 abort(); 292 } 293 } 294 295 /* 296 * We are done decoding Base-64 chars. Let's see if we ended 297 * on a byte boundary, and/or with erroneous trailing characters. 298 */ 299 300 if (ch == Pad64) { /*%< We got a pad char. */ 301 ch = *src++; /*%< Skip it, get next. */ 302 switch (state) { 303 case 0: /*%< Invalid = in first position */ 304 case 1: /*%< Invalid = in second position */ 305 return -1; 306 307 case 2: /*%< Valid, means one byte of info */ 308 /* Skip any number of spaces. */ 309 for (; ch != '\0'; ch = (u_char) *src++) 310 if (!isspace(ch)) 311 break; 312 /* Make sure there is another trailing = sign. */ 313 if (ch != Pad64) 314 return -1; 315 ch = *src++; /*%< Skip the = */ 316 /* Fall through to "single trailing =" case. */ 317 /* FALLTHROUGH */ 318 319 case 3: /*%< Valid, means two bytes of info */ 320 /* 321 * We know this char is an =. Is there anything but 322 * whitespace after it? 323 */ 324 for (; ch != '\0'; ch = (u_char) *src++) 325 if (!isspace(ch)) 326 return -1; 327 328 /* 329 * Now make sure for cases 2 and 3 that the "extra" 330 * bits that slopped past the last full byte were 331 * zeros. If we don't check them, they become a 332 * subliminal channel. 333 */ 334 if (target && tarindex < targsize && 335 target[tarindex] != 0) 336 return -1; 337 } 338 } else { 339 /* 340 * We ended by seeing the end of the string. Make sure we 341 * have no partial bytes lying around. 342 */ 343 if (state != 0) 344 return -1; 345 } 346 347 assert(tarindex <= (size_t)INT_MAX); 348 return (int)tarindex; 349 } 350 351 /*! \file */ 352