xref: /freebsd/lib/libc/net/base64.c (revision f0adf7f5cdd241db2f2c817683191a6ef64a4e95)
1 /*
2  * Copyright (c) 1996, 1998 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/cdefs.h>
44 __FBSDID("$FreeBSD$");
45 
46 #include <sys/types.h>
47 #include <sys/param.h>
48 #include <sys/socket.h>
49 
50 #include <netinet/in.h>
51 #include <arpa/inet.h>
52 #include <arpa/nameser.h>
53 
54 #include <ctype.h>
55 #include <resolv.h>
56 #include <stdio.h>
57 #include <stdlib.h>
58 #include <string.h>
59 
60 #define Assert(Cond) if (!(Cond)) abort()
61 
62 static const char Base64[] =
63 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
64 static const char Pad64 = '=';
65 
66 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
67    The following encoding technique is taken from RFC 1521 by Borenstein
68    and Freed.  It is reproduced here in a slightly edited form for
69    convenience.
70 
71    A 65-character subset of US-ASCII is used, enabling 6 bits to be
72    represented per printable character. (The extra 65th character, "=",
73    is used to signify a special processing function.)
74 
75    The encoding process represents 24-bit groups of input bits as output
76    strings of 4 encoded characters. Proceeding from left to right, a
77    24-bit input group is formed by concatenating 3 8-bit input groups.
78    These 24 bits are then treated as 4 concatenated 6-bit groups, each
79    of which is translated into a single digit in the base64 alphabet.
80 
81    Each 6-bit group is used as an index into an array of 64 printable
82    characters. The character referenced by the index is placed in the
83    output string.
84 
85                          Table 1: The Base64 Alphabet
86 
87       Value Encoding  Value Encoding  Value Encoding  Value Encoding
88           0 A            17 R            34 i            51 z
89           1 B            18 S            35 j            52 0
90           2 C            19 T            36 k            53 1
91           3 D            20 U            37 l            54 2
92           4 E            21 V            38 m            55 3
93           5 F            22 W            39 n            56 4
94           6 G            23 X            40 o            57 5
95           7 H            24 Y            41 p            58 6
96           8 I            25 Z            42 q            59 7
97           9 J            26 a            43 r            60 8
98          10 K            27 b            44 s            61 9
99          11 L            28 c            45 t            62 +
100          12 M            29 d            46 u            63 /
101          13 N            30 e            47 v
102          14 O            31 f            48 w         (pad) =
103          15 P            32 g            49 x
104          16 Q            33 h            50 y
105 
106    Special processing is performed if fewer than 24 bits are available
107    at the end of the data being encoded.  A full encoding quantum is
108    always completed at the end of a quantity.  When fewer than 24 input
109    bits are available in an input group, zero bits are added (on the
110    right) to form an integral number of 6-bit groups.  Padding at the
111    end of the data is performed using the '=' character.
112 
113    Since all base64 input is an integral number of octets, only the
114          -------------------------------------------------
115    following cases can arise:
116 
117        (1) the final quantum of encoding input is an integral
118            multiple of 24 bits; here, the final unit of encoded
119 	   output will be an integral multiple of 4 characters
120 	   with no "=" padding,
121        (2) the final quantum of encoding input is exactly 8 bits;
122            here, the final unit of encoded output will be two
123 	   characters followed by two "=" padding characters, or
124        (3) the final quantum of encoding input is exactly 16 bits;
125            here, the final unit of encoded output will be three
126 	   characters followed by one "=" padding character.
127    */
128 
129 int
130 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
131 	size_t datalength = 0;
132 	u_char input[3];
133 	u_char output[4];
134 	size_t i;
135 
136 	while (2 < srclength) {
137 		input[0] = *src++;
138 		input[1] = *src++;
139 		input[2] = *src++;
140 		srclength -= 3;
141 
142 		output[0] = input[0] >> 2;
143 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
144 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
145 		output[3] = input[2] & 0x3f;
146 		Assert(output[0] < 64);
147 		Assert(output[1] < 64);
148 		Assert(output[2] < 64);
149 		Assert(output[3] < 64);
150 
151 		if (datalength + 4 > targsize)
152 			return (-1);
153 		target[datalength++] = Base64[output[0]];
154 		target[datalength++] = Base64[output[1]];
155 		target[datalength++] = Base64[output[2]];
156 		target[datalength++] = Base64[output[3]];
157 	}
158 
159 	/* Now we worry about padding. */
160 	if (0 != srclength) {
161 		/* Get what's left. */
162 		input[0] = input[1] = input[2] = '\0';
163 		for (i = 0; i < srclength; i++)
164 			input[i] = *src++;
165 
166 		output[0] = input[0] >> 2;
167 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
168 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
169 		Assert(output[0] < 64);
170 		Assert(output[1] < 64);
171 		Assert(output[2] < 64);
172 
173 		if (datalength + 4 > targsize)
174 			return (-1);
175 		target[datalength++] = Base64[output[0]];
176 		target[datalength++] = Base64[output[1]];
177 		if (srclength == 1)
178 			target[datalength++] = Pad64;
179 		else
180 			target[datalength++] = Base64[output[2]];
181 		target[datalength++] = Pad64;
182 	}
183 	if (datalength >= targsize)
184 		return (-1);
185 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
186 	return (datalength);
187 }
188 
189 /* skips all whitespace anywhere.
190    converts characters, four at a time, starting at (or after)
191    src from base - 64 numbers into three 8 bit bytes in the target area.
192    it returns the number of data bytes stored at the target, or -1 on error.
193  */
194 
195 int
196 b64_pton(src, target, targsize)
197 	char const *src;
198 	u_char *target;
199 	size_t targsize;
200 {
201 	int tarindex, state, ch;
202 	char *pos;
203 
204 	state = 0;
205 	tarindex = 0;
206 
207 	while ((ch = *src++) != '\0') {
208 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
209 			continue;
210 
211 		if (ch == Pad64)
212 			break;
213 
214 		pos = strchr(Base64, ch);
215 		if (pos == 0) 		/* A non-base64 character. */
216 			return (-1);
217 
218 		switch (state) {
219 		case 0:
220 			if (target) {
221 				if ((size_t)tarindex >= targsize)
222 					return (-1);
223 				target[tarindex] = (pos - Base64) << 2;
224 			}
225 			state = 1;
226 			break;
227 		case 1:
228 			if (target) {
229 				if ((size_t)tarindex + 1 >= targsize)
230 					return (-1);
231 				target[tarindex]   |=  (pos - Base64) >> 4;
232 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
233 							<< 4 ;
234 			}
235 			tarindex++;
236 			state = 2;
237 			break;
238 		case 2:
239 			if (target) {
240 				if ((size_t)tarindex + 1 >= targsize)
241 					return (-1);
242 				target[tarindex]   |=  (pos - Base64) >> 2;
243 				target[tarindex+1]  = ((pos - Base64) & 0x03)
244 							<< 6;
245 			}
246 			tarindex++;
247 			state = 3;
248 			break;
249 		case 3:
250 			if (target) {
251 				if ((size_t)tarindex >= targsize)
252 					return (-1);
253 				target[tarindex] |= (pos - Base64);
254 			}
255 			tarindex++;
256 			state = 0;
257 			break;
258 		default:
259 			abort();
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 = *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 ((void)NULL; ch != '\0'; ch = *src++)
278 				if (!isspace((unsigned char)ch))
279 					break;
280 			/* Make sure there is another trailing = sign. */
281 			if (ch != Pad64)
282 				return (-1);
283 			ch = *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 ((void)NULL; ch != '\0'; ch = *src++)
293 				if (!isspace((unsigned char)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 && target[tarindex] != 0)
303 				return (-1);
304 		}
305 	} else {
306 		/*
307 		 * We ended by seeing the end of the string.  Make sure we
308 		 * have no partial bytes lying around.
309 		 */
310 		if (state != 0)
311 			return (-1);
312 	}
313 
314 	return (tarindex);
315 }
316