xref: /freebsd/lib/libc/net/base64.c (revision 61afd5bb22d787b0641523e7b9b95c964d669bd5)
1 /*
2  * Copyright (c) 1996 by Internet Software Consortium.
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
9  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
10  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
11  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
12  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
13  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
14  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
15  * SOFTWARE.
16  */
17 
18 /*
19  * Portions Copyright (c) 1995 by International Business Machines, Inc.
20  *
21  * International Business Machines, Inc. (hereinafter called IBM) grants
22  * permission under its copyrights to use, copy, modify, and distribute this
23  * Software with or without fee, provided that the above copyright notice and
24  * all paragraphs of this notice appear in all copies, and that the name of IBM
25  * not be used in connection with the marketing of any product incorporating
26  * the Software or modifications thereof, without specific, written prior
27  * permission.
28  *
29  * To the extent it has a right to do so, IBM grants an immunity from suit
30  * under its patents, if any, for the use, sale or manufacture of products to
31  * the extent that such products are used for performing Domain Name System
32  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
33  * granted for any product per se or for any other function of any product.
34  *
35  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
36  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
37  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
38  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
39  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
40  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
41  */
42 
43 #include <sys/types.h>
44 #include <sys/param.h>
45 #include <sys/socket.h>
46 #include <netinet/in.h>
47 #include <arpa/inet.h>
48 #include <arpa/nameser.h>
49 
50 #include <ctype.h>
51 #include <resolv.h>
52 #include <stdio.h>
53 
54 #if defined(BSD) && (BSD >= 199103) && defined(AF_INET6)
55 # include <stdlib.h>
56 # include <string.h>
57 #else
58 # include "../conf/portability.h"
59 #endif
60 
61 #define Assert(Cond) if (!(Cond)) abort()
62 
63 static const char Base64[] =
64 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
65 static const char Pad64 = '=';
66 
67 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
68    The following encoding technique is taken from RFC 1521 by Borenstein
69    and Freed.  It is reproduced here in a slightly edited form for
70    convenience.
71 
72    A 65-character subset of US-ASCII is used, enabling 6 bits to be
73    represented per printable character. (The extra 65th character, "=",
74    is used to signify a special processing function.)
75 
76    The encoding process represents 24-bit groups of input bits as output
77    strings of 4 encoded characters. Proceeding from left to right, a
78    24-bit input group is formed by concatenating 3 8-bit input groups.
79    These 24 bits are then treated as 4 concatenated 6-bit groups, each
80    of which is translated into a single digit in the base64 alphabet.
81 
82    Each 6-bit group is used as an index into an array of 64 printable
83    characters. The character referenced by the index is placed in the
84    output string.
85 
86                          Table 1: The Base64 Alphabet
87 
88       Value Encoding  Value Encoding  Value Encoding  Value Encoding
89           0 A            17 R            34 i            51 z
90           1 B            18 S            35 j            52 0
91           2 C            19 T            36 k            53 1
92           3 D            20 U            37 l            54 2
93           4 E            21 V            38 m            55 3
94           5 F            22 W            39 n            56 4
95           6 G            23 X            40 o            57 5
96           7 H            24 Y            41 p            58 6
97           8 I            25 Z            42 q            59 7
98           9 J            26 a            43 r            60 8
99          10 K            27 b            44 s            61 9
100          11 L            28 c            45 t            62 +
101          12 M            29 d            46 u            63 /
102          13 N            30 e            47 v
103          14 O            31 f            48 w         (pad) =
104          15 P            32 g            49 x
105          16 Q            33 h            50 y
106 
107    Special processing is performed if fewer than 24 bits are available
108    at the end of the data being encoded.  A full encoding quantum is
109    always completed at the end of a quantity.  When fewer than 24 input
110    bits are available in an input group, zero bits are added (on the
111    right) to form an integral number of 6-bit groups.  Padding at the
112    end of the data is performed using the '=' character.
113 
114    Since all base64 input is an integral number of octets, only the
115          -------------------------------------------------
116    following cases can arise:
117 
118        (1) the final quantum of encoding input is an integral
119            multiple of 24 bits; here, the final unit of encoded
120 	   output will be an integral multiple of 4 characters
121 	   with no "=" padding,
122        (2) the final quantum of encoding input is exactly 8 bits;
123            here, the final unit of encoded output will be two
124 	   characters followed by two "=" padding characters, or
125        (3) the final quantum of encoding input is exactly 16 bits;
126            here, the final unit of encoded output will be three
127 	   characters followed by one "=" padding character.
128    */
129 
130 int
131 b64_ntop(src, srclength, target, targsize)
132 	u_char const *src;
133 	size_t srclength;
134 	char *target;
135 	size_t targsize;
136 {
137 	size_t datalength = 0;
138 	u_char input[3];
139 	u_char output[4];
140 	int i;
141 
142 	while (2 < srclength) {
143 		input[0] = *src++;
144 		input[1] = *src++;
145 		input[2] = *src++;
146 		srclength -= 3;
147 
148 		output[0] = input[0] >> 2;
149 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
150 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
151 		output[3] = input[2] & 0x3f;
152 		Assert(output[0] < 64);
153 		Assert(output[1] < 64);
154 		Assert(output[2] < 64);
155 		Assert(output[3] < 64);
156 
157 		if (datalength + 4 > targsize)
158 			return (-1);
159 		target[datalength++] = Base64[output[0]];
160 		target[datalength++] = Base64[output[1]];
161 		target[datalength++] = Base64[output[2]];
162 		target[datalength++] = Base64[output[3]];
163 	}
164 
165 	/* Now we worry about padding. */
166 	if (0 != srclength) {
167 		/* Get what's left. */
168 		input[0] = input[1] = input[2] = '\0';
169 		for (i = 0; i < srclength; i++)
170 			input[i] = *src++;
171 
172 		output[0] = input[0] >> 2;
173 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
174 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
175 		Assert(output[0] < 64);
176 		Assert(output[1] < 64);
177 		Assert(output[2] < 64);
178 
179 		if (datalength + 4 > targsize)
180 			return (-1);
181 		target[datalength++] = Base64[output[0]];
182 		target[datalength++] = Base64[output[1]];
183 		if (srclength == 1)
184 			target[datalength++] = Pad64;
185 		else
186 			target[datalength++] = Base64[output[2]];
187 		target[datalength++] = Pad64;
188 	}
189 	if (datalength >= targsize)
190 		return (-1);
191 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
192 	return (datalength);
193 }
194 
195 /* skips all whitespace anywhere.
196    converts characters, four at a time, starting at (or after)
197    src from base - 64 numbers into three 8 bit bytes in the target area.
198    it returns the number of data bytes stored at the target, or -1 on error.
199  */
200 
201 int
202 b64_pton(src, target, targsize)
203 	char const *src;
204 	u_char *target;
205 	size_t targsize;
206 {
207 	int tarindex, state, ch;
208 	char *pos;
209 
210 	state = 0;
211 	tarindex = 0;
212 
213 	while ((ch = *src++) != '\0') {
214 		if (isspace(ch))	/* Skip whitespace anywhere. */
215 			continue;
216 
217 		if (ch == Pad64)
218 			break;
219 
220 		pos = strchr(Base64, ch);
221 		if (pos == 0) 		/* A non-base64 character. */
222 			return (-1);
223 
224 		switch (state) {
225 		case 0:
226 			if (target) {
227 				if (tarindex >= targsize)
228 					return (-1);
229 				target[tarindex] = (pos - Base64) << 2;
230 			}
231 			state = 1;
232 			break;
233 		case 1:
234 			if (target) {
235 				if (tarindex + 1 >= targsize)
236 					return (-1);
237 				target[tarindex]   |=  (pos - Base64) >> 4;
238 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
239 							<< 4 ;
240 			}
241 			tarindex++;
242 			state = 2;
243 			break;
244 		case 2:
245 			if (target) {
246 				if (tarindex + 1 >= targsize)
247 					return (-1);
248 				target[tarindex]   |=  (pos - Base64) >> 2;
249 				target[tarindex+1]  = ((pos - Base64) & 0x03)
250 							<< 6;
251 			}
252 			tarindex++;
253 			state = 3;
254 			break;
255 		case 3:
256 			if (target) {
257 				if (tarindex >= targsize)
258 					return (-1);
259 				target[tarindex] |= (pos - Base64);
260 			}
261 			tarindex++;
262 			state = 0;
263 			break;
264 		default:
265 			abort();
266 		}
267 	}
268 
269 	/*
270 	 * We are done decoding Base-64 chars.  Let's see if we ended
271 	 * on a byte boundary, and/or with erroneous trailing characters.
272 	 */
273 
274 	if (ch == Pad64) {		/* We got a pad char. */
275 		ch = *src++;		/* Skip it, get next. */
276 		switch (state) {
277 		case 0:		/* Invalid = in first position */
278 		case 1:		/* Invalid = in second position */
279 			return (-1);
280 
281 		case 2:		/* Valid, means one byte of info */
282 			/* Skip any number of spaces. */
283 			for (NULL; ch != '\0'; ch = *src++)
284 				if (!isspace(ch))
285 					break;
286 			/* Make sure there is another trailing = sign. */
287 			if (ch != Pad64)
288 				return (-1);
289 			ch = *src++;		/* Skip the = */
290 			/* Fall through to "single trailing =" case. */
291 			/* FALLTHROUGH */
292 
293 		case 3:		/* Valid, means two bytes of info */
294 			/*
295 			 * We know this char is an =.  Is there anything but
296 			 * whitespace after it?
297 			 */
298 			for (NULL; ch != '\0'; ch = *src++)
299 				if (!isspace(ch))
300 					return (-1);
301 
302 			/*
303 			 * Now make sure for cases 2 and 3 that the "extra"
304 			 * bits that slopped past the last full byte were
305 			 * zeros.  If we don't check them, they become a
306 			 * subliminal channel.
307 			 */
308 			if (target && target[tarindex] != 0)
309 				return (-1);
310 		}
311 	} else {
312 		/*
313 		 * We ended by seeing the end of the string.  Make sure we
314 		 * have no partial bytes lying around.
315 		 */
316 		if (state != 0)
317 			return (-1);
318 	}
319 
320 	return (tarindex);
321 }
322