xref: /openbsd-src/lib/libc/net/base64.c (revision 50b7afb2c2c0993b0894d4e34bf857cb13ed9c80)
1 /*	$OpenBSD: base64.c,v 1.7 2013/12/31 02:32:56 tedu 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/types.h>
46 #include <sys/param.h>
47 #include <sys/socket.h>
48 #include <netinet/in.h>
49 #include <arpa/inet.h>
50 #include <arpa/nameser.h>
51 
52 #include <ctype.h>
53 #include <resolv.h>
54 #include <stdio.h>
55 
56 #include <stdlib.h>
57 #include <string.h>
58 
59 static const char Base64[] =
60 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
61 static const char Pad64 = '=';
62 
63 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
64    The following encoding technique is taken from RFC 1521 by Borenstein
65    and Freed.  It is reproduced here in a slightly edited form for
66    convenience.
67 
68    A 65-character subset of US-ASCII is used, enabling 6 bits to be
69    represented per printable character. (The extra 65th character, "=",
70    is used to signify a special processing function.)
71 
72    The encoding process represents 24-bit groups of input bits as output
73    strings of 4 encoded characters. Proceeding from left to right, a
74    24-bit input group is formed by concatenating 3 8-bit input groups.
75    These 24 bits are then treated as 4 concatenated 6-bit groups, each
76    of which is translated into a single digit in the base64 alphabet.
77 
78    Each 6-bit group is used as an index into an array of 64 printable
79    characters. The character referenced by the index is placed in the
80    output string.
81 
82                          Table 1: The Base64 Alphabet
83 
84       Value Encoding  Value Encoding  Value Encoding  Value Encoding
85           0 A            17 R            34 i            51 z
86           1 B            18 S            35 j            52 0
87           2 C            19 T            36 k            53 1
88           3 D            20 U            37 l            54 2
89           4 E            21 V            38 m            55 3
90           5 F            22 W            39 n            56 4
91           6 G            23 X            40 o            57 5
92           7 H            24 Y            41 p            58 6
93           8 I            25 Z            42 q            59 7
94           9 J            26 a            43 r            60 8
95          10 K            27 b            44 s            61 9
96          11 L            28 c            45 t            62 +
97          12 M            29 d            46 u            63 /
98          13 N            30 e            47 v
99          14 O            31 f            48 w         (pad) =
100          15 P            32 g            49 x
101          16 Q            33 h            50 y
102 
103    Special processing is performed if fewer than 24 bits are available
104    at the end of the data being encoded.  A full encoding quantum is
105    always completed at the end of a quantity.  When fewer than 24 input
106    bits are available in an input group, zero bits are added (on the
107    right) to form an integral number of 6-bit groups.  Padding at the
108    end of the data is performed using the '=' character.
109 
110    Since all base64 input is an integral number of octets, only the
111          -------------------------------------------------
112    following cases can arise:
113 
114        (1) the final quantum of encoding input is an integral
115            multiple of 24 bits; here, the final unit of encoded
116 	   output will be an integral multiple of 4 characters
117 	   with no "=" padding,
118        (2) the final quantum of encoding input is exactly 8 bits;
119            here, the final unit of encoded output will be two
120 	   characters followed by two "=" padding characters, or
121        (3) the final quantum of encoding input is exactly 16 bits;
122            here, the final unit of encoded output will be three
123 	   characters followed by one "=" padding character.
124    */
125 
126 int
127 b64_ntop(src, srclength, target, targsize)
128 	u_char const *src;
129 	size_t srclength;
130 	char *target;
131 	size_t targsize;
132 {
133 	size_t datalength = 0;
134 	u_char input[3];
135 	u_char output[4];
136 	int i;
137 
138 	while (2 < srclength) {
139 		input[0] = *src++;
140 		input[1] = *src++;
141 		input[2] = *src++;
142 		srclength -= 3;
143 
144 		output[0] = input[0] >> 2;
145 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
146 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
147 		output[3] = input[2] & 0x3f;
148 
149 		if (datalength + 4 > targsize)
150 			return (-1);
151 		target[datalength++] = Base64[output[0]];
152 		target[datalength++] = Base64[output[1]];
153 		target[datalength++] = Base64[output[2]];
154 		target[datalength++] = Base64[output[3]];
155 	}
156 
157 	/* Now we worry about padding. */
158 	if (0 != srclength) {
159 		/* Get what's left. */
160 		input[0] = input[1] = input[2] = '\0';
161 		for (i = 0; i < srclength; i++)
162 			input[i] = *src++;
163 
164 		output[0] = input[0] >> 2;
165 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
166 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
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 b64_pton(src, target, targsize)
192 	char const *src;
193 	u_char *target;
194 	size_t targsize;
195 {
196 	int tarindex, state, ch;
197 	u_char nextbyte;
198 	char *pos;
199 
200 	state = 0;
201 	tarindex = 0;
202 
203 	while ((ch = (unsigned char)*src++) != '\0') {
204 		if (isspace(ch))	/* Skip whitespace anywhere. */
205 			continue;
206 
207 		if (ch == Pad64)
208 			break;
209 
210 		pos = strchr(Base64, ch);
211 		if (pos == 0) 		/* A non-base64 character. */
212 			return (-1);
213 
214 		switch (state) {
215 		case 0:
216 			if (target) {
217 				if (tarindex >= targsize)
218 					return (-1);
219 				target[tarindex] = (pos - Base64) << 2;
220 			}
221 			state = 1;
222 			break;
223 		case 1:
224 			if (target) {
225 				if (tarindex >= targsize)
226 					return (-1);
227 				target[tarindex]   |=  (pos - Base64) >> 4;
228 				nextbyte = ((pos - Base64) & 0x0f) << 4;
229 				if (tarindex + 1 < targsize)
230 					target[tarindex+1] = nextbyte;
231 				else if (nextbyte)
232 					return (-1);
233 			}
234 			tarindex++;
235 			state = 2;
236 			break;
237 		case 2:
238 			if (target) {
239 				if (tarindex >= targsize)
240 					return (-1);
241 				target[tarindex]   |=  (pos - Base64) >> 2;
242 				nextbyte = ((pos - Base64) & 0x03) << 6;
243 				if (tarindex + 1 < targsize)
244 					target[tarindex+1] = nextbyte;
245 				else if (nextbyte)
246 					return (-1);
247 			}
248 			tarindex++;
249 			state = 3;
250 			break;
251 		case 3:
252 			if (target) {
253 				if (tarindex >= targsize)
254 					return (-1);
255 				target[tarindex] |= (pos - Base64);
256 			}
257 			tarindex++;
258 			state = 0;
259 			break;
260 		}
261 	}
262 
263 	/*
264 	 * We are done decoding Base-64 chars.  Let's see if we ended
265 	 * on a byte boundary, and/or with erroneous trailing characters.
266 	 */
267 
268 	if (ch == Pad64) {			/* We got a pad char. */
269 		ch = (unsigned char)*src++;	/* Skip it, get next. */
270 		switch (state) {
271 		case 0:		/* Invalid = in first position */
272 		case 1:		/* Invalid = in second position */
273 			return (-1);
274 
275 		case 2:		/* Valid, means one byte of info */
276 			/* Skip any number of spaces. */
277 			for (; ch != '\0'; ch = (unsigned char)*src++)
278 				if (!isspace(ch))
279 					break;
280 			/* Make sure there is another trailing = sign. */
281 			if (ch != Pad64)
282 				return (-1);
283 			ch = (unsigned char)*src++;		/* Skip the = */
284 			/* Fall through to "single trailing =" case. */
285 			/* FALLTHROUGH */
286 
287 		case 3:		/* Valid, means two bytes of info */
288 			/*
289 			 * We know this char is an =.  Is there anything but
290 			 * whitespace after it?
291 			 */
292 			for (; ch != '\0'; ch = (unsigned char)*src++)
293 				if (!isspace(ch))
294 					return (-1);
295 
296 			/*
297 			 * Now make sure for cases 2 and 3 that the "extra"
298 			 * bits that slopped past the last full byte were
299 			 * zeros.  If we don't check them, they become a
300 			 * subliminal channel.
301 			 */
302 			if (target && tarindex < targsize &&
303 			    target[tarindex] != 0)
304 				return (-1);
305 		}
306 	} else {
307 		/*
308 		 * We ended by seeing the end of the string.  Make sure we
309 		 * have no partial bytes lying around.
310 		 */
311 		if (state != 0)
312 			return (-1);
313 	}
314 
315 	return (tarindex);
316 }
317