lib/base64.c

   1 #include "gsasl.h"
   2
   3 /* This file is copied from GNU libc - libc/resolv/base64.c */
   4
   5 /*
   6  * Copyright (c) 1996-1999 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 <ctype.h>
  48 #include <stdio.h>
  49 #include <stdlib.h>
  50 #include <string.h>
  51
  52 #define Assert(Cond) if (!(Cond)) abort()
  53
  54 static const char Base64[] =
  55         "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
  56 static const char Pad64 = '=';
  57
  58 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
  59    The following encoding technique is taken from RFC 1521 by Borenstein
  60    and Freed.  It is reproduced here in a slightly edited form for
  61    convenience.
  62
  63    A 65-character subset of US-ASCII is used, enabling 6 bits to be
  64    represented per printable character. (The extra 65th character, "=",
  65    is used to signify a special processing function.)
  66
  67    The encoding process represents 24-bit groups of input bits as output
  68    strings of 4 encoded characters. Proceeding from left to right, a
  69    24-bit input group is formed by concatenating 3 8-bit input groups.
  70    These 24 bits are then treated as 4 concatenated 6-bit groups, each
  71    of which is translated into a single digit in the base64 alphabet.
  72
  73    Each 6-bit group is used as an index into an array of 64 printable
  74    characters. The character referenced by the index is placed in the
  75    output string.
  76
  77                          Table 1: The Base64 Alphabet
  78
  79       Value Encoding  Value Encoding  Value Encoding  Value Encoding
  80           0 A            17 R            34 i            51 z
  81           1 B            18 S            35 j            52 0
  82           2 C            19 T            36 k            53 1
  83           3 D            20 U            37 l            54 2
  84           4 E            21 V            38 m            55 3
  85           5 F            22 W            39 n            56 4
  86           6 G            23 X            40 o            57 5
  87           7 H            24 Y            41 p            58 6
  88           8 I            25 Z            42 q            59 7
  89           9 J            26 a            43 r            60 8
  90          10 K            27 b            44 s            61 9
  91          11 L            28 c            45 t            62 +
  92          12 M            29 d            46 u            63 /
  93          13 N            30 e            47 v
  94          14 O            31 f            48 w         (pad) =
  95          15 P            32 g            49 x
  96          16 Q            33 h            50 y
  97
  98    Special processing is performed if fewer than 24 bits are available
  99    at the end of the data being encoded.  A full encoding quantum is
 100    always completed at the end of a quantity.  When fewer than 24 input
 101    bits are available in an input group, zero bits are added (on the
 102    right) to form an integral number of 6-bit groups.  Padding at the
 103    end of the data is performed using the '=' character.
 104
 105    Since all base64 input is an integral number of octets, only the
 106          -------------------------------------------------
 107    following cases can arise:
 108
 109        (1) the final quantum of encoding input is an integral
 110            multiple of 24 bits; here, the final unit of encoded
 111            output will be an integral multiple of 4 characters
 112            with no "=" padding,
 113        (2) the final quantum of encoding input is exactly 8 bits;
 114            here, the final unit of encoded output will be two
 115            characters followed by two "=" padding characters, or
 116        (3) the final quantum of encoding input is exactly 16 bits;
 117            here, the final unit of encoded output will be three
 118            characters followed by one "=" padding character.
 119    */
 120
 121 /**
 122  * gsasl_base64_encode:
 123  * @src: input byte array
 124  * @srclength: size of input byte array
 125  * @target: output byte array
 126  * @targsize: size of output byte array
 127  *
 128  * Encode data as base64.  Converts characters, three at a time,
 129  * starting at src into four base64 characters in the target area
 130  * until the entire input buffer is encoded.
 131  *
 132  * Return value: Returns the number of data bytes stored at the
 133  * target, or -1 on error.
 134  **/
 135 int
 136 gsasl_base64_encode (unsigned char const *src,
 137                      size_t srclength,
 138                      char *target,
 139                      size_t targsize)
 140 {
 141         size_t datalength = 0;
 142         unsigned char input[3];
 143         unsigned char output[4];
 144         size_t i;
 145
 146         while (2 < srclength) {
 147                 input[0] = *src++;
 148                 input[1] = *src++;
 149                 input[2] = *src++;
 150                 srclength -= 3;
 151
 152                 output[0] = input[0] >> 2;
 153                 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
 154                 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
 155                 output[3] = input[2] & 0x3f;
 156                 Assert(output[0] < 64);
 157                 Assert(output[1] < 64);
 158                 Assert(output[2] < 64);
 159                 Assert(output[3] < 64);
 160
 161                 if (datalength + 4 > targsize)
 162                         return (-1);
 163                 target[datalength++] = Base64[output[0]];
 164                 target[datalength++] = Base64[output[1]];
 165                 target[datalength++] = Base64[output[2]];
 166                 target[datalength++] = Base64[output[3]];
 167         }
 168
 169         /* Now we worry about padding. */
 170         if (0 != srclength) {
 171                 /* Get what's left. */
 172                 input[0] = input[1] = input[2] = '\0';
 173                 for (i = 0; i < srclength; i++)
 174                         input[i] = *src++;
 175
 176                 output[0] = input[0] >> 2;
 177                 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
 178                 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
 179                 Assert(output[0] < 64);
 180                 Assert(output[1] < 64);
 181                 Assert(output[2] < 64);
 182
 183                 if (datalength + 4 > targsize)
 184                         return (-1);
 185                 target[datalength++] = Base64[output[0]];
 186                 target[datalength++] = Base64[output[1]];
 187                 if (srclength == 1)
 188                         target[datalength++] = Pad64;
 189                 else
 190                         target[datalength++] = Base64[output[2]];
 191                 target[datalength++] = Pad64;
 192         }
 193         if (datalength >= targsize)
 194                 return (-1);
 195         target[datalength] = '\0';      /* Returned value doesn't count \0. */
 196         return (datalength);
 197 }
 198
 199 /**
 200  * gsasl_base64_decode:
 201  * @src: input byte array
 202  * @target: output byte array
 203  * @targsize: size of output byte array
 204  *
 205  * Decode Base64 data.  Skips all whitespace anywhere.  Converts
 206  * characters, four at a time, starting at (or after) src from Base64
 207  * numbers into three 8 bit bytes in the target area.
 208  *
 209  * Return value: Returns the number of data bytes stored at the
 210  * target, or -1 on error.
 211  **/
 212 int
 213 gsasl_base64_decode (char const *src,
 214                      unsigned char *target,
 215                      size_t targsize)
 216 {
 217         int tarindex, state, ch;
 218         char *pos;
 219
 220         state = 0;
 221         tarindex = 0;
 222
 223         while ((ch = *src++) != '\0') {
 224                 if (isspace(ch))        /* Skip whitespace anywhere. */
 225                         continue;
 226
 227                 if (ch == Pad64)
 228                         break;
 229
 230                 pos = strchr(Base64, ch);
 231                 if (pos == 0)           /* A non-base64 character. */
 232                         return (-1);
 233
 234                 switch (state) {
 235                 case 0:
 236                         if (target) {
 237                                 if ((size_t)tarindex >= targsize)
 238                                         return (-1);
 239                                 target[tarindex] = (pos - Base64) << 2;
 240                         }
 241                         state = 1;
 242                         break;
 243                 case 1:
 244                         if (target) {
 245                                 if ((size_t)tarindex + 1 >= targsize)
 246                                         return (-1);
 247                                 target[tarindex]   |=  (pos - Base64) >> 4;
 248                                 target[tarindex+1]  = ((pos - Base64) & 0x0f)
 249                                                         << 4 ;
 250                         }
 251                         tarindex++;
 252                         state = 2;
 253                         break;
 254                 case 2:
 255                         if (target) {
 256                                 if ((size_t)tarindex + 1 >= targsize)
 257                                         return (-1);
 258                                 target[tarindex]   |=  (pos - Base64) >> 2;
 259                                 target[tarindex+1]  = ((pos - Base64) & 0x03)
 260                                                         << 6;
 261                         }
 262                         tarindex++;
 263                         state = 3;
 264                         break;
 265                 case 3:
 266                         if (target) {
 267                                 if ((size_t)tarindex >= targsize)
 268                                         return (-1);
 269                                 target[tarindex] |= (pos - Base64);
 270                         }
 271                         tarindex++;
 272                         state = 0;
 273                         break;
 274                 default:
 275                         abort();
 276                 }
 277         }
 278
 279         /*
 280          * We are done decoding Base-64 chars.  Let's see if we ended
 281          * on a byte boundary, and/or with erroneous trailing characters.
 282          */
 283
 284         if (ch == Pad64) {              /* We got a pad char. */
 285                 ch = *src++;            /* Skip it, get next. */
 286                 switch (state) {
 287                 case 0:         /* Invalid = in first position */
 288                 case 1:         /* Invalid = in second position */
 289                         return (-1);
 290
 291                 case 2:         /* Valid, means one byte of info */
 292                         /* Skip any number of spaces. */
 293                         for ((void)NULL; ch != '\0'; ch = *src++)
 294                                 if (!isspace(ch))
 295                                         break;
 296                         /* Make sure there is another trailing = sign. */
 297                         if (ch != Pad64)
 298                                 return (-1);
 299                         ch = *src++;            /* Skip the = */
 300                         /* Fall through to "single trailing =" case. */
 301                         /* FALLTHROUGH */
 302
 303                 case 3:         /* Valid, means two bytes of info */
 304                         /*
 305                          * We know this char is an =.  Is there anything but
 306                          * whitespace after it?
 307                          */
 308                         for ((void)NULL; ch != '\0'; ch = *src++)
 309                                 if (!isspace(ch))
 310                                         return (-1);
 311
 312                         /*
 313                          * Now make sure for cases 2 and 3 that the "extra"
 314                          * bits that slopped past the last full byte were
 315                          * zeros.  If we don't check them, they become a
 316                          * subliminal channel.
 317                          */
 318                         if (target && target[tarindex] != 0)
 319                                 return (-1);
 320                 }
 321         } else {
 322                 /*
 323                  * We ended by seeing the end of the string.  Make sure we
 324                  * have no partial bytes lying around.
 325                  */
 326                 if (state != 0)
 327                         return (-1);
 328         }
 329
 330         return (tarindex);
 331 }