xref: /freebsd/lib/libc/net/base64.c (revision 315ee00f)
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 #include <sys/param.h>
45 #include <sys/socket.h>
46 
47 #include <netinet/in.h>
48 #include <arpa/inet.h>
49 #include <arpa/nameser.h>
50 
51 #include <ctype.h>
52 #include <resolv.h>
53 #include <stdio.h>
54 #include <stdlib.h>
55 #include <string.h>
56 
57 #define Assert(Cond) if (!(Cond)) abort()
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(u_char const *src, size_t srclength, char *target, size_t targsize) {
128 	size_t datalength = 0;
129 	u_char input[3];
130 	u_char output[4];
131 	size_t i;
132 
133 	while (2 < srclength) {
134 		input[0] = *src++;
135 		input[1] = *src++;
136 		input[2] = *src++;
137 		srclength -= 3;
138 
139 		output[0] = input[0] >> 2;
140 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
141 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
142 		output[3] = input[2] & 0x3f;
143 		Assert(output[0] < 64);
144 		Assert(output[1] < 64);
145 		Assert(output[2] < 64);
146 		Assert(output[3] < 64);
147 
148 		if (datalength + 4 > targsize)
149 			return (-1);
150 		target[datalength++] = Base64[output[0]];
151 		target[datalength++] = Base64[output[1]];
152 		target[datalength++] = Base64[output[2]];
153 		target[datalength++] = Base64[output[3]];
154 	}
155 
156 	/* Now we worry about padding. */
157 	if (0 != srclength) {
158 		/* Get what's left. */
159 		input[0] = input[1] = input[2] = '\0';
160 		for (i = 0; i < srclength; i++)
161 			input[i] = *src++;
162 
163 		output[0] = input[0] >> 2;
164 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
165 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
166 		Assert(output[0] < 64);
167 		Assert(output[1] < 64);
168 		Assert(output[2] < 64);
169 
170 		if (datalength + 4 > targsize)
171 			return (-1);
172 		target[datalength++] = Base64[output[0]];
173 		target[datalength++] = Base64[output[1]];
174 		if (srclength == 1)
175 			target[datalength++] = Pad64;
176 		else
177 			target[datalength++] = Base64[output[2]];
178 		target[datalength++] = Pad64;
179 	}
180 	if (datalength >= targsize)
181 		return (-1);
182 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
183 	return (datalength);
184 }
185 
186 /* skips all whitespace anywhere.
187    converts characters, four at a time, starting at (or after)
188    src from base - 64 numbers into three 8 bit bytes in the target area.
189    it returns the number of data bytes stored at the target, or -1 on error.
190  */
191 
192 int
193 b64_pton(const char *src, u_char *target, size_t targsize)
194 {
195 	int tarindex, state, ch;
196 	u_char nextbyte;
197 	char *pos;
198 
199 	state = 0;
200 	tarindex = 0;
201 
202 	while ((ch = *src++) != '\0') {
203 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
204 			continue;
205 
206 		if (ch == Pad64)
207 			break;
208 
209 		pos = strchr(Base64, ch);
210 		if (pos == NULL)		/* 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 >= targsize)
225 					return (-1);
226 				target[tarindex]   |=  (pos - Base64) >> 4;
227 				nextbyte = ((pos - Base64) & 0x0f) << 4;
228 				if ((size_t)tarindex + 1 < targsize)
229 					target[tarindex + 1] = nextbyte;
230 				else if (nextbyte)
231 					return (-1);
232 			}
233 			tarindex++;
234 			state = 2;
235 			break;
236 		case 2:
237 			if (target) {
238 				if ((size_t)tarindex >= targsize)
239 					return (-1);
240 				target[tarindex]   |=  (pos - Base64) >> 2;
241 				nextbyte = ((pos - Base64) & 0x03) << 6;
242 				if ((size_t)tarindex + 1 < targsize)
243 					target[tarindex + 1] = nextbyte;
244 				else if (nextbyte)
245 					return (-1);
246 			}
247 			tarindex++;
248 			state = 3;
249 			break;
250 		case 3:
251 			if (target) {
252 				if ((size_t)tarindex >= targsize)
253 					return (-1);
254 				target[tarindex] |= (pos - Base64);
255 			}
256 			tarindex++;
257 			state = 0;
258 			break;
259 		default:
260 			abort();
261 		}
262 	}
263 
264 	/*
265 	 * We are done decoding Base-64 chars.  Let's see if we ended
266 	 * on a byte boundary, and/or with erroneous trailing characters.
267 	 */
268 
269 	if (ch == Pad64) {		/* We got a pad char. */
270 		ch = *src++;		/* Skip it, get next. */
271 		switch (state) {
272 		case 0:		/* Invalid = in first position */
273 		case 1:		/* Invalid = in second position */
274 			return (-1);
275 
276 		case 2:		/* Valid, means one byte of info */
277 			/* Skip any number of spaces. */
278 			for ((void)NULL; ch != '\0'; ch = *src++)
279 				if (!isspace((unsigned char)ch))
280 					break;
281 			/* Make sure there is another trailing = sign. */
282 			if (ch != Pad64)
283 				return (-1);
284 			ch = *src++;		/* Skip the = */
285 			/* Fall through to "single trailing =" case. */
286 			/* FALLTHROUGH */
287 
288 		case 3:		/* Valid, means two bytes of info */
289 			/*
290 			 * We know this char is an =.  Is there anything but
291 			 * whitespace after it?
292 			 */
293 			for ((void)NULL; ch != '\0'; ch = *src++)
294 				if (!isspace((unsigned char)ch))
295 					return (-1);
296 
297 			/*
298 			 * Now make sure for cases 2 and 3 that the "extra"
299 			 * bits that slopped past the last full byte were
300 			 * zeros.  If we don't check them, they become a
301 			 * subliminal channel.
302 			 */
303 			if (target && (size_t)tarindex < targsize &&
304 			    target[tarindex] != 0)
305 				return (-1);
306 		}
307 	} else {
308 		/*
309 		 * We ended by seeing the end of the string.  Make sure we
310 		 * have no partial bytes lying around.
311 		 */
312 		if (state != 0)
313 			return (-1);
314 	}
315 
316 	return (tarindex);
317 }
318