xref: /netbsd-src/lib/libc/net/base64.c (revision 2a399c6883d870daece976daec6ffa7bb7f934ce)
1 /*	$NetBSD: base64.c,v 1.3 1997/07/13 19:57:30 christos Exp $	*/
2 
3 /*
4  * Copyright (c) 1996 by Internet Software Consortium.
5  *
6  * Permission to use, copy, modify, and distribute this software for any
7  * purpose with or without fee is hereby granted, provided that the above
8  * copyright notice and this permission notice appear in all copies.
9  *
10  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
11  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
12  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
13  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
14  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
15  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
16  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
17  * 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 __RCSID("$NetBSD: base64.c,v 1.3 1997/07/13 19:57:30 christos Exp $");
48 #endif /* LIBC_SCCS and not lint */
49 
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 #include <arpa/nameser.h>
57 
58 #include <ctype.h>
59 #include <resolv.h>
60 #include <stdio.h>
61 
62 #if defined(BSD) && (BSD >= 199103) && defined(AF_INET6)
63 # include <stdlib.h>
64 # include <string.h>
65 #else
66 # include "../conf/portability.h"
67 #endif
68 
69 #define Assert(Cond) if (!(Cond)) abort()
70 
71 static const char Base64[] =
72 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
73 static const char Pad64 = '=';
74 
75 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
76    The following encoding technique is taken from RFC 1521 by Borenstein
77    and Freed.  It is reproduced here in a slightly edited form for
78    convenience.
79 
80    A 65-character subset of US-ASCII is used, enabling 6 bits to be
81    represented per printable character. (The extra 65th character, "=",
82    is used to signify a special processing function.)
83 
84    The encoding process represents 24-bit groups of input bits as output
85    strings of 4 encoded characters. Proceeding from left to right, a
86    24-bit input group is formed by concatenating 3 8-bit input groups.
87    These 24 bits are then treated as 4 concatenated 6-bit groups, each
88    of which is translated into a single digit in the base64 alphabet.
89 
90    Each 6-bit group is used as an index into an array of 64 printable
91    characters. The character referenced by the index is placed in the
92    output string.
93 
94                          Table 1: The Base64 Alphabet
95 
96       Value Encoding  Value Encoding  Value Encoding  Value Encoding
97           0 A            17 R            34 i            51 z
98           1 B            18 S            35 j            52 0
99           2 C            19 T            36 k            53 1
100           3 D            20 U            37 l            54 2
101           4 E            21 V            38 m            55 3
102           5 F            22 W            39 n            56 4
103           6 G            23 X            40 o            57 5
104           7 H            24 Y            41 p            58 6
105           8 I            25 Z            42 q            59 7
106           9 J            26 a            43 r            60 8
107          10 K            27 b            44 s            61 9
108          11 L            28 c            45 t            62 +
109          12 M            29 d            46 u            63 /
110          13 N            30 e            47 v
111          14 O            31 f            48 w         (pad) =
112          15 P            32 g            49 x
113          16 Q            33 h            50 y
114 
115    Special processing is performed if fewer than 24 bits are available
116    at the end of the data being encoded.  A full encoding quantum is
117    always completed at the end of a quantity.  When fewer than 24 input
118    bits are available in an input group, zero bits are added (on the
119    right) to form an integral number of 6-bit groups.  Padding at the
120    end of the data is performed using the '=' character.
121 
122    Since all base64 input is an integral number of octets, only the
123          -------------------------------------------------
124    following cases can arise:
125 
126        (1) the final quantum of encoding input is an integral
127            multiple of 24 bits; here, the final unit of encoded
128 	   output will be an integral multiple of 4 characters
129 	   with no "=" padding,
130        (2) the final quantum of encoding input is exactly 8 bits;
131            here, the final unit of encoded output will be two
132 	   characters followed by two "=" padding characters, or
133        (3) the final quantum of encoding input is exactly 16 bits;
134            here, the final unit of encoded output will be three
135 	   characters followed by one "=" padding character.
136    */
137 
138 int
139 b64_ntop(src, srclength, target, targsize)
140 	u_char const *src;
141 	size_t srclength;
142 	char *target;
143 	size_t targsize;
144 {
145 	size_t datalength = 0;
146 	u_char input[3];
147 	u_char output[4];
148 	int i;
149 
150 	while (2 < srclength) {
151 		input[0] = *src++;
152 		input[1] = *src++;
153 		input[2] = *src++;
154 		srclength -= 3;
155 
156 		output[0] = input[0] >> 2;
157 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
158 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
159 		output[3] = input[2] & 0x3f;
160 		Assert(output[0] < 64);
161 		Assert(output[1] < 64);
162 		Assert(output[2] < 64);
163 		Assert(output[3] < 64);
164 
165 		if (datalength + 4 > targsize)
166 			return (-1);
167 		target[datalength++] = Base64[output[0]];
168 		target[datalength++] = Base64[output[1]];
169 		target[datalength++] = Base64[output[2]];
170 		target[datalength++] = Base64[output[3]];
171 	}
172 
173 	/* Now we worry about padding. */
174 	if (0 != srclength) {
175 		/* Get what's left. */
176 		input[0] = input[1] = input[2] = '\0';
177 		for (i = 0; i < srclength; i++)
178 			input[i] = *src++;
179 
180 		output[0] = input[0] >> 2;
181 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
182 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
183 		Assert(output[0] < 64);
184 		Assert(output[1] < 64);
185 		Assert(output[2] < 64);
186 
187 		if (datalength + 4 > targsize)
188 			return (-1);
189 		target[datalength++] = Base64[output[0]];
190 		target[datalength++] = Base64[output[1]];
191 		if (srclength == 1)
192 			target[datalength++] = Pad64;
193 		else
194 			target[datalength++] = Base64[output[2]];
195 		target[datalength++] = Pad64;
196 	}
197 	if (datalength >= targsize)
198 		return (-1);
199 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
200 	return (datalength);
201 }
202 
203 /* skips all whitespace anywhere.
204    converts characters, four at a time, starting at (or after)
205    src from base - 64 numbers into three 8 bit bytes in the target area.
206    it returns the number of data bytes stored at the target, or -1 on error.
207  */
208 
209 int
210 b64_pton(src, target, targsize)
211 	char const *src;
212 	u_char *target;
213 	size_t targsize;
214 {
215 	int tarindex, state, ch;
216 	char *pos;
217 
218 	state = 0;
219 	tarindex = 0;
220 
221 	while ((ch = *src++) != '\0') {
222 		if (isspace(ch))	/* Skip whitespace anywhere. */
223 			continue;
224 
225 		if (ch == Pad64)
226 			break;
227 
228 		pos = strchr(Base64, ch);
229 		if (pos == 0) 		/* A non-base64 character. */
230 			return (-1);
231 
232 		switch (state) {
233 		case 0:
234 			if (target) {
235 				if (tarindex >= targsize)
236 					return (-1);
237 				target[tarindex] = (pos - Base64) << 2;
238 			}
239 			state = 1;
240 			break;
241 		case 1:
242 			if (target) {
243 				if (tarindex + 1 >= targsize)
244 					return (-1);
245 				target[tarindex]   |=  (pos - Base64) >> 4;
246 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
247 							<< 4 ;
248 			}
249 			tarindex++;
250 			state = 2;
251 			break;
252 		case 2:
253 			if (target) {
254 				if (tarindex + 1 >= targsize)
255 					return (-1);
256 				target[tarindex]   |=  (pos - Base64) >> 2;
257 				target[tarindex+1]  = ((pos - Base64) & 0x03)
258 							<< 6;
259 			}
260 			tarindex++;
261 			state = 3;
262 			break;
263 		case 3:
264 			if (target) {
265 				if (tarindex >= targsize)
266 					return (-1);
267 				target[tarindex] |= (pos - Base64);
268 			}
269 			tarindex++;
270 			state = 0;
271 			break;
272 		default:
273 			abort();
274 		}
275 	}
276 
277 	/*
278 	 * We are done decoding Base-64 chars.  Let's see if we ended
279 	 * on a byte boundary, and/or with erroneous trailing characters.
280 	 */
281 
282 	if (ch == Pad64) {		/* We got a pad char. */
283 		ch = *src++;		/* Skip it, get next. */
284 		switch (state) {
285 		case 0:		/* Invalid = in first position */
286 		case 1:		/* Invalid = in second position */
287 			return (-1);
288 
289 		case 2:		/* Valid, means one byte of info */
290 			/* Skip any number of spaces. */
291 			for (; ch != '\0'; ch = *src++)
292 				if (!isspace(ch))
293 					break;
294 			/* Make sure there is another trailing = sign. */
295 			if (ch != Pad64)
296 				return (-1);
297 			ch = *src++;		/* Skip the = */
298 			/* Fall through to "single trailing =" case. */
299 			/* FALLTHROUGH */
300 
301 		case 3:		/* Valid, means two bytes of info */
302 			/*
303 			 * We know this char is an =.  Is there anything but
304 			 * whitespace after it?
305 			 */
306 			for (; ch != '\0'; ch = *src++)
307 				if (!isspace(ch))
308 					return (-1);
309 
310 			/*
311 			 * Now make sure for cases 2 and 3 that the "extra"
312 			 * bits that slopped past the last full byte were
313 			 * zeros.  If we don't check them, they become a
314 			 * subliminal channel.
315 			 */
316 			if (target && target[tarindex] != 0)
317 				return (-1);
318 		}
319 	} else {
320 		/*
321 		 * We ended by seeing the end of the string.  Make sure we
322 		 * have no partial bytes lying around.
323 		 */
324 		if (state != 0)
325 			return (-1);
326 	}
327 
328 	return (tarindex);
329 }
330