xref: /netbsd-src/external/bsd/openldap/dist/libraries/liblutil/base64.c (revision 946379e7b37692fc43f68eb0d1c10daa0a7f3b6c)
1 /*	$NetBSD: base64.c,v 1.1.1.4 2014/05/28 09:58:45 tron Exp $	*/
2 
3 /* base64.c -- routines to encode/decode base64 data */
4 /* $OpenLDAP$ */
5 /* This work is part of OpenLDAP Software <http://www.openldap.org/>.
6  *
7  * Copyright 1998-2014 The OpenLDAP Foundation.
8  * Portions Copyright 1998-2003 Kurt D. Zeilenga.
9  * Portions Copyright 1995 IBM Corporation.
10  * All rights reserved.
11  *
12  * Redistribution and use in source and binary forms, with or without
13  * modification, are permitted only as authorized by the OpenLDAP
14  * Public License.
15  *
16  * A copy of this license is available in the file LICENSE in the
17  * top-level directory of the distribution or, alternatively, at
18  * <http://www.OpenLDAP.org/license.html>.
19  */
20 /* Portions Copyright (c) 1996, 1998 by Internet Software Consortium.
21  *
22  * Permission to use, copy, modify, and distribute this software for any
23  * purpose with or without fee is hereby granted, provided that the above
24  * copyright notice and this permission notice appear in all copies.
25  *
26  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
27  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
28  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
29  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
30  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
31  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
32  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
33  * SOFTWARE.
34  */
35 /* This work is based upon Base64 routines (developed by IBM) found
36  * Berkeley Internet Name Daemon (BIND) as distributed by ISC.  They
37  * were adapted for inclusion in OpenLDAP Software by Kurt D. Zeilenga.
38  */
39 
40 #include "portable.h"
41 
42 #include <ac/assert.h>
43 #include <ac/stdlib.h>
44 #include <ac/ctype.h>
45 #include <ac/string.h>
46 
47 /* include socket.h to get sys/types.h and/or winsock2.h */
48 #include <ac/socket.h>
49 
50 #include "lutil.h"
51 
52 static const char Base64[] =
53 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
54 static const char Pad64 = '=';
55 
56 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
57    The following encoding technique is taken from RFC 1521 by Borenstein
58    and Freed.  It is reproduced here in a slightly edited form for
59    convenience.
60 
61    A 65-character subset of US-ASCII is used, enabling 6 bits to be
62    represented per printable character. (The extra 65th character, "=",
63    is used to signify a special processing function.)
64 
65    The encoding process represents 24-bit groups of input bits as output
66    strings of 4 encoded characters. Proceeding from left to right, a
67    24-bit input group is formed by concatenating 3 8-bit input groups.
68    These 24 bits are then treated as 4 concatenated 6-bit groups, each
69    of which is translated into a single digit in the base64 alphabet.
70 
71    Each 6-bit group is used as an index into an array of 64 printable
72    characters. The character referenced by the index is placed in the
73    output string.
74 
75                          Table 1: The Base64 Alphabet
76 
77       Value Encoding  Value Encoding  Value Encoding  Value Encoding
78           0 A            17 R            34 i            51 z
79           1 B            18 S            35 j            52 0
80           2 C            19 T            36 k            53 1
81           3 D            20 U            37 l            54 2
82           4 E            21 V            38 m            55 3
83           5 F            22 W            39 n            56 4
84           6 G            23 X            40 o            57 5
85           7 H            24 Y            41 p            58 6
86           8 I            25 Z            42 q            59 7
87           9 J            26 a            43 r            60 8
88          10 K            27 b            44 s            61 9
89          11 L            28 c            45 t            62 +
90          12 M            29 d            46 u            63 /
91          13 N            30 e            47 v
92          14 O            31 f            48 w         (pad) =
93          15 P            32 g            49 x
94          16 Q            33 h            50 y
95 
96    Special processing is performed if fewer than 24 bits are available
97    at the end of the data being encoded.  A full encoding quantum is
98    always completed at the end of a quantity.  When fewer than 24 input
99    bits are available in an input group, zero bits are added (on the
100    right) to form an integral number of 6-bit groups.  Padding at the
101    end of the data is performed using the '=' character.
102 
103    Since all base64 input is an integral number of octets, only the
104          -------------------------------------------------
105    following cases can arise:
106 
107        (1) the final quantum of encoding input is an integral
108            multiple of 24 bits; here, the final unit of encoded
109 	   output will be an integral multiple of 4 characters
110 	   with no "=" padding,
111        (2) the final quantum of encoding input is exactly 8 bits;
112            here, the final unit of encoded output will be two
113 	   characters followed by two "=" padding characters, or
114        (3) the final quantum of encoding input is exactly 16 bits;
115            here, the final unit of encoded output will be three
116 	   characters followed by one "=" padding character.
117    */
118 
119 int
120 lutil_b64_ntop(
121 	u_char const *src,
122 	size_t srclength,
123 	char *target,
124 	size_t targsize)
125 {
126 	size_t datalength = 0;
127 	u_char input[3];
128 	u_char output[4];
129 	size_t i;
130 
131 	while (2 < srclength) {
132 		input[0] = *src++;
133 		input[1] = *src++;
134 		input[2] = *src++;
135 		srclength -= 3;
136 
137 		output[0] = input[0] >> 2;
138 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
139 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
140 		output[3] = input[2] & 0x3f;
141 		assert(output[0] < 64);
142 		assert(output[1] < 64);
143 		assert(output[2] < 64);
144 		assert(output[3] < 64);
145 
146 		if (datalength + 4 > targsize)
147 			return (-1);
148 		target[datalength++] = Base64[output[0]];
149 		target[datalength++] = Base64[output[1]];
150 		target[datalength++] = Base64[output[2]];
151 		target[datalength++] = Base64[output[3]];
152 	}
153 
154 	/* Now we worry about padding. */
155 	if (0 != srclength) {
156 		/* Get what's left. */
157 		input[0] = input[1] = input[2] = '\0';
158 		for (i = 0; i < srclength; i++)
159 			input[i] = *src++;
160 
161 		output[0] = input[0] >> 2;
162 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
163 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
164 		assert(output[0] < 64);
165 		assert(output[1] < 64);
166 		assert(output[2] < 64);
167 
168 		if (datalength + 4 > targsize)
169 			return (-1);
170 		target[datalength++] = Base64[output[0]];
171 		target[datalength++] = Base64[output[1]];
172 		if (srclength == 1)
173 			target[datalength++] = Pad64;
174 		else
175 			target[datalength++] = Base64[output[2]];
176 		target[datalength++] = Pad64;
177 	}
178 	if (datalength >= targsize)
179 		return (-1);
180 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
181 	return (datalength);
182 }
183 
184 /* skips all whitespace anywhere.
185    converts characters, four at a time, starting at (or after)
186    src from base - 64 numbers into three 8 bit bytes in the target area.
187    it returns the number of data bytes stored at the target, or -1 on error.
188  */
189 
190 int
191 lutil_b64_pton(
192 	char const *src,
193 	u_char *target,
194 	size_t targsize)
195 {
196 	int tarindex, state, ch;
197 	char *pos;
198 
199 	state = 0;
200 	tarindex = 0;
201 
202 	while ((ch = *src++) != '\0') {
203 		if (isascii(ch) && isspace(ch))	/* Skip whitespace anywhere. */
204 			continue;
205 
206 		if (ch == Pad64)
207 			break;
208 
209 		pos = strchr(Base64, ch);
210 		if (pos == 0) 		/* A non-base64 character. */
211 			return (-1);
212 
213 		switch (state) {
214 		case 0:
215 			if (target) {
216 				if ((size_t)tarindex >= targsize)
217 					return (-1);
218 				target[tarindex] = (pos - Base64) << 2;
219 			}
220 			state = 1;
221 			break;
222 		case 1:
223 			if (target) {
224 				if ((size_t)tarindex + 1 >= targsize)
225 					return (-1);
226 				target[tarindex]   |=  (pos - Base64) >> 4;
227 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
228 							<< 4 ;
229 			}
230 			tarindex++;
231 			state = 2;
232 			break;
233 		case 2:
234 			if (target) {
235 				if ((size_t)tarindex + 1 >= targsize)
236 					return (-1);
237 				target[tarindex]   |=  (pos - Base64) >> 2;
238 				target[tarindex+1]  = ((pos - Base64) & 0x03)
239 							<< 6;
240 			}
241 			tarindex++;
242 			state = 3;
243 			break;
244 		case 3:
245 			if (target) {
246 				if ((size_t)tarindex >= targsize)
247 					return (-1);
248 				target[tarindex] |= (pos - Base64);
249 			}
250 			tarindex++;
251 			state = 0;
252 			break;
253 		default:
254 			abort();
255 		}
256 	}
257 
258 	/*
259 	 * We are done decoding Base-64 chars.  Let's see if we ended
260 	 * on a byte boundary, and/or with erroneous trailing characters.
261 	 */
262 
263 	if (ch == Pad64) {		/* We got a pad char. */
264 		ch = *src++;		/* Skip it, get next. */
265 		switch (state) {
266 		case 0:		/* Invalid = in first position */
267 		case 1:		/* Invalid = in second position */
268 			return (-1);
269 
270 		case 2:		/* Valid, means one byte of info */
271 			/* Skip any number of spaces. */
272 			for ((void)NULL; ch != '\0'; ch = *src++)
273 				if (! (isascii(ch) && isspace(ch)))
274 					break;
275 			/* Make sure there is another trailing = sign. */
276 			if (ch != Pad64)
277 				return (-1);
278 			ch = *src++;		/* Skip the = */
279 			/* Fall through to "single trailing =" case. */
280 			/* FALLTHROUGH */
281 
282 		case 3:		/* Valid, means two bytes of info */
283 			/*
284 			 * We know this char is an =.  Is there anything but
285 			 * whitespace after it?
286 			 */
287 			for ((void)NULL; ch != '\0'; ch = *src++)
288 				if (! (isascii(ch) && isspace(ch)))
289 					return (-1);
290 
291 			/*
292 			 * Now make sure for cases 2 and 3 that the "extra"
293 			 * bits that slopped past the last full byte were
294 			 * zeros.  If we don't check them, they become a
295 			 * subliminal channel.
296 			 */
297 			if (target && target[tarindex] != 0)
298 				return (-1);
299 		}
300 	} else {
301 		/*
302 		 * We ended by seeing the end of the string.  Make sure we
303 		 * have no partial bytes lying around.
304 		 */
305 		if (state != 0)
306 			return (-1);
307 	}
308 
309 	return (tarindex);
310 }
311