1 /* base64.c -- routines to encode/decode base64 data */
3 /* This work is part of OpenLDAP Software <http://www.openldap.org/>.
5 * Copyright 1998-2016 The OpenLDAP Foundation.
6 * Portions Copyright 1998-2003 Kurt D. Zeilenga.
7 * Portions Copyright 1995 IBM Corporation.
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted only as authorized by the OpenLDAP
14 * A copy of this license is available in the file LICENSE in the
15 * top-level directory of the distribution or, alternatively, at
16 * <http://www.OpenLDAP.org/license.html>.
18 /* Portions Copyright (c) 1996, 1998 by Internet Software Consortium.
20 * Permission to use, copy, modify, and distribute this software for any
21 * purpose with or without fee is hereby granted, provided that the above
22 * copyright notice and this permission notice appear in all copies.
24 * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
25 * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
26 * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
27 * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
28 * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
29 * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
30 * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
33 /* This work is based upon Base64 routines (developed by IBM) found
34 * Berkeley Internet Name Daemon (BIND) as distributed by ISC. They
35 * were adapted for inclusion in OpenLDAP Software by Kurt D. Zeilenga.
40 #include <ac/assert.h>
41 #include <ac/stdlib.h>
43 #include <ac/string.h>
45 /* include socket.h to get sys/types.h and/or winsock2.h */
46 #include <ac/socket.h>
50 static const char Base64[] =
51 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
52 static const char Pad64 = '=';
54 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
55 The following encoding technique is taken from RFC 1521 by Borenstein
56 and Freed. It is reproduced here in a slightly edited form for
59 A 65-character subset of US-ASCII is used, enabling 6 bits to be
60 represented per printable character. (The extra 65th character, "=",
61 is used to signify a special processing function.)
63 The encoding process represents 24-bit groups of input bits as output
64 strings of 4 encoded characters. Proceeding from left to right, a
65 24-bit input group is formed by concatenating 3 8-bit input groups.
66 These 24 bits are then treated as 4 concatenated 6-bit groups, each
67 of which is translated into a single digit in the base64 alphabet.
69 Each 6-bit group is used as an index into an array of 64 printable
70 characters. The character referenced by the index is placed in the
73 Table 1: The Base64 Alphabet
75 Value Encoding Value Encoding Value Encoding Value Encoding
90 14 O 31 f 48 w (pad) =
94 Special processing is performed if fewer than 24 bits are available
95 at the end of the data being encoded. A full encoding quantum is
96 always completed at the end of a quantity. When fewer than 24 input
97 bits are available in an input group, zero bits are added (on the
98 right) to form an integral number of 6-bit groups. Padding at the
99 end of the data is performed using the '=' character.
101 Since all base64 input is an integral number of octets, only the
102 -------------------------------------------------
103 following cases can arise:
105 (1) the final quantum of encoding input is an integral
106 multiple of 24 bits; here, the final unit of encoded
107 output will be an integral multiple of 4 characters
109 (2) the final quantum of encoding input is exactly 8 bits;
110 here, the final unit of encoded output will be two
111 characters followed by two "=" padding characters, or
112 (3) the final quantum of encoding input is exactly 16 bits;
113 here, the final unit of encoded output will be three
114 characters followed by one "=" padding character.
124 size_t datalength = 0;
129 while (2 < srclength) {
135 output[0] = input[0] >> 2;
136 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
137 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
138 output[3] = input[2] & 0x3f;
139 assert(output[0] < 64);
140 assert(output[1] < 64);
141 assert(output[2] < 64);
142 assert(output[3] < 64);
144 if (datalength + 4 > targsize)
146 target[datalength++] = Base64[output[0]];
147 target[datalength++] = Base64[output[1]];
148 target[datalength++] = Base64[output[2]];
149 target[datalength++] = Base64[output[3]];
152 /* Now we worry about padding. */
153 if (0 != srclength) {
154 /* Get what's left. */
155 input[0] = input[1] = input[2] = '\0';
156 for (i = 0; i < srclength; i++)
159 output[0] = input[0] >> 2;
160 output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
161 output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
162 assert(output[0] < 64);
163 assert(output[1] < 64);
164 assert(output[2] < 64);
166 if (datalength + 4 > targsize)
168 target[datalength++] = Base64[output[0]];
169 target[datalength++] = Base64[output[1]];
171 target[datalength++] = Pad64;
173 target[datalength++] = Base64[output[2]];
174 target[datalength++] = Pad64;
176 if (datalength >= targsize)
178 target[datalength] = '\0'; /* Returned value doesn't count \0. */
182 /* skips all whitespace anywhere.
183 converts characters, four at a time, starting at (or after)
184 src from base - 64 numbers into three 8 bit bytes in the target area.
185 it returns the number of data bytes stored at the target, or -1 on error.
194 int tarindex, state, ch;
200 while ((ch = *src++) != '\0') {
201 if (isascii(ch) && isspace(ch)) /* Skip whitespace anywhere. */
207 pos = strchr(Base64, ch);
208 if (pos == 0) /* A non-base64 character. */
214 if ((size_t)tarindex >= targsize)
216 target[tarindex] = (pos - Base64) << 2;
222 if ((size_t)tarindex + 1 >= targsize)
224 target[tarindex] |= (pos - Base64) >> 4;
225 target[tarindex+1] = ((pos - Base64) & 0x0f)
233 if ((size_t)tarindex + 1 >= targsize)
235 target[tarindex] |= (pos - Base64) >> 2;
236 target[tarindex+1] = ((pos - Base64) & 0x03)
244 if ((size_t)tarindex >= targsize)
246 target[tarindex] |= (pos - Base64);
257 * We are done decoding Base-64 chars. Let's see if we ended
258 * on a byte boundary, and/or with erroneous trailing characters.
261 if (ch == Pad64) { /* We got a pad char. */
262 ch = *src++; /* Skip it, get next. */
264 case 0: /* Invalid = in first position */
265 case 1: /* Invalid = in second position */
268 case 2: /* Valid, means one byte of info */
269 /* Skip any number of spaces. */
270 for ((void)NULL; ch != '\0'; ch = *src++)
271 if (! (isascii(ch) && isspace(ch)))
273 /* Make sure there is another trailing = sign. */
276 ch = *src++; /* Skip the = */
277 /* Fall through to "single trailing =" case. */
280 case 3: /* Valid, means two bytes of info */
282 * We know this char is an =. Is there anything but
283 * whitespace after it?
285 for ((void)NULL; ch != '\0'; ch = *src++)
286 if (! (isascii(ch) && isspace(ch)))
290 * Now make sure for cases 2 and 3 that the "extra"
291 * bits that slopped past the last full byte were
292 * zeros. If we don't check them, they become a
293 * subliminal channel.
295 if (target && target[tarindex] != 0)
300 * We ended by seeing the end of the string. Make sure we
301 * have no partial bytes lying around.