xref: /freebsd/lib/libc/net/base64.c (revision 42249ef2)
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/param.h>
47 #include <sys/socket.h>
48 
49 #include <netinet/in.h>
50 #include <arpa/inet.h>
51 #include <arpa/nameser.h>
52 
53 #include <ctype.h>
54 #include <resolv.h>
55 #include <stdio.h>
56 #include <stdlib.h>
57 #include <string.h>
58 
59 #define Assert(Cond) if (!(Cond)) abort()
60 
61 static const char Base64[] =
62 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
63 static const char Pad64 = '=';
64 
65 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
66    The following encoding technique is taken from RFC 1521 by Borenstein
67    and Freed.  It is reproduced here in a slightly edited form for
68    convenience.
69 
70    A 65-character subset of US-ASCII is used, enabling 6 bits to be
71    represented per printable character. (The extra 65th character, "=",
72    is used to signify a special processing function.)
73 
74    The encoding process represents 24-bit groups of input bits as output
75    strings of 4 encoded characters. Proceeding from left to right, a
76    24-bit input group is formed by concatenating 3 8-bit input groups.
77    These 24 bits are then treated as 4 concatenated 6-bit groups, each
78    of which is translated into a single digit in the base64 alphabet.
79 
80    Each 6-bit group is used as an index into an array of 64 printable
81    characters. The character referenced by the index is placed in the
82    output string.
83 
84                          Table 1: The Base64 Alphabet
85 
86       Value Encoding  Value Encoding  Value Encoding  Value Encoding
87           0 A            17 R            34 i            51 z
88           1 B            18 S            35 j            52 0
89           2 C            19 T            36 k            53 1
90           3 D            20 U            37 l            54 2
91           4 E            21 V            38 m            55 3
92           5 F            22 W            39 n            56 4
93           6 G            23 X            40 o            57 5
94           7 H            24 Y            41 p            58 6
95           8 I            25 Z            42 q            59 7
96           9 J            26 a            43 r            60 8
97          10 K            27 b            44 s            61 9
98          11 L            28 c            45 t            62 +
99          12 M            29 d            46 u            63 /
100          13 N            30 e            47 v
101          14 O            31 f            48 w         (pad) =
102          15 P            32 g            49 x
103          16 Q            33 h            50 y
104 
105    Special processing is performed if fewer than 24 bits are available
106    at the end of the data being encoded.  A full encoding quantum is
107    always completed at the end of a quantity.  When fewer than 24 input
108    bits are available in an input group, zero bits are added (on the
109    right) to form an integral number of 6-bit groups.  Padding at the
110    end of the data is performed using the '=' character.
111 
112    Since all base64 input is an integral number of octets, only the
113          -------------------------------------------------
114    following cases can arise:
115 
116        (1) the final quantum of encoding input is an integral
117            multiple of 24 bits; here, the final unit of encoded
118 	   output will be an integral multiple of 4 characters
119 	   with no "=" padding,
120        (2) the final quantum of encoding input is exactly 8 bits;
121            here, the final unit of encoded output will be two
122 	   characters followed by two "=" padding characters, or
123        (3) the final quantum of encoding input is exactly 16 bits;
124            here, the final unit of encoded output will be three
125 	   characters followed by one "=" padding character.
126    */
127 
128 int
129 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
130 	size_t datalength = 0;
131 	u_char input[3];
132 	u_char output[4];
133 	size_t i;
134 
135 	while (2 < srclength) {
136 		input[0] = *src++;
137 		input[1] = *src++;
138 		input[2] = *src++;
139 		srclength -= 3;
140 
141 		output[0] = input[0] >> 2;
142 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
143 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
144 		output[3] = input[2] & 0x3f;
145 		Assert(output[0] < 64);
146 		Assert(output[1] < 64);
147 		Assert(output[2] < 64);
148 		Assert(output[3] < 64);
149 
150 		if (datalength + 4 > targsize)
151 			return (-1);
152 		target[datalength++] = Base64[output[0]];
153 		target[datalength++] = Base64[output[1]];
154 		target[datalength++] = Base64[output[2]];
155 		target[datalength++] = Base64[output[3]];
156 	}
157 
158 	/* Now we worry about padding. */
159 	if (0 != srclength) {
160 		/* Get what's left. */
161 		input[0] = input[1] = input[2] = '\0';
162 		for (i = 0; i < srclength; i++)
163 			input[i] = *src++;
164 
165 		output[0] = input[0] >> 2;
166 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
167 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
168 		Assert(output[0] < 64);
169 		Assert(output[1] < 64);
170 		Assert(output[2] < 64);
171 
172 		if (datalength + 4 > targsize)
173 			return (-1);
174 		target[datalength++] = Base64[output[0]];
175 		target[datalength++] = Base64[output[1]];
176 		if (srclength == 1)
177 			target[datalength++] = Pad64;
178 		else
179 			target[datalength++] = Base64[output[2]];
180 		target[datalength++] = Pad64;
181 	}
182 	if (datalength >= targsize)
183 		return (-1);
184 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
185 	return (datalength);
186 }
187 
188 /* skips all whitespace anywhere.
189    converts characters, four at a time, starting at (or after)
190    src from base - 64 numbers into three 8 bit bytes in the target area.
191    it returns the number of data bytes stored at the target, or -1 on error.
192  */
193 
194 int
195 b64_pton(const char *src, u_char *target, size_t targsize)
196 {
197 	int tarindex, state, ch;
198 	u_char nextbyte;
199 	char *pos;
200 
201 	state = 0;
202 	tarindex = 0;
203 
204 	while ((ch = *src++) != '\0') {
205 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
206 			continue;
207 
208 		if (ch == Pad64)
209 			break;
210 
211 		pos = strchr(Base64, ch);
212 		if (pos == NULL)		/* A non-base64 character. */
213 			return (-1);
214 
215 		switch (state) {
216 		case 0:
217 			if (target) {
218 				if ((size_t)tarindex >= targsize)
219 					return (-1);
220 				target[tarindex] = (pos - Base64) << 2;
221 			}
222 			state = 1;
223 			break;
224 		case 1:
225 			if (target) {
226 				if ((size_t)tarindex >= targsize)
227 					return (-1);
228 				target[tarindex]   |=  (pos - Base64) >> 4;
229 				nextbyte = ((pos - Base64) & 0x0f) << 4;
230 				if ((size_t)tarindex + 1 < targsize)
231 					target[tarindex + 1] = nextbyte;
232 				else if (nextbyte)
233 					return (-1);
234 			}
235 			tarindex++;
236 			state = 2;
237 			break;
238 		case 2:
239 			if (target) {
240 				if ((size_t)tarindex >= targsize)
241 					return (-1);
242 				target[tarindex]   |=  (pos - Base64) >> 2;
243 				nextbyte = ((pos - Base64) & 0x03) << 6;
244 				if ((size_t)tarindex + 1 < targsize)
245 					target[tarindex + 1] = nextbyte;
246 				else if (nextbyte)
247 					return (-1);
248 			}
249 			tarindex++;
250 			state = 3;
251 			break;
252 		case 3:
253 			if (target) {
254 				if ((size_t)tarindex >= targsize)
255 					return (-1);
256 				target[tarindex] |= (pos - Base64);
257 			}
258 			tarindex++;
259 			state = 0;
260 			break;
261 		default:
262 			abort();
263 		}
264 	}
265 
266 	/*
267 	 * We are done decoding Base-64 chars.  Let's see if we ended
268 	 * on a byte boundary, and/or with erroneous trailing characters.
269 	 */
270 
271 	if (ch == Pad64) {		/* We got a pad char. */
272 		ch = *src++;		/* Skip it, get next. */
273 		switch (state) {
274 		case 0:		/* Invalid = in first position */
275 		case 1:		/* Invalid = in second position */
276 			return (-1);
277 
278 		case 2:		/* Valid, means one byte of info */
279 			/* Skip any number of spaces. */
280 			for ((void)NULL; ch != '\0'; ch = *src++)
281 				if (!isspace((unsigned char)ch))
282 					break;
283 			/* Make sure there is another trailing = sign. */
284 			if (ch != Pad64)
285 				return (-1);
286 			ch = *src++;		/* Skip the = */
287 			/* Fall through to "single trailing =" case. */
288 			/* FALLTHROUGH */
289 
290 		case 3:		/* Valid, means two bytes of info */
291 			/*
292 			 * We know this char is an =.  Is there anything but
293 			 * whitespace after it?
294 			 */
295 			for ((void)NULL; ch != '\0'; ch = *src++)
296 				if (!isspace((unsigned char)ch))
297 					return (-1);
298 
299 			/*
300 			 * Now make sure for cases 2 and 3 that the "extra"
301 			 * bits that slopped past the last full byte were
302 			 * zeros.  If we don't check them, they become a
303 			 * subliminal channel.
304 			 */
305 			if (target && (size_t)tarindex < targsize &&
306 			    target[tarindex] != 0)
307 				return (-1);
308 		}
309 	} else {
310 		/*
311 		 * We ended by seeing the end of the string.  Make sure we
312 		 * have no partial bytes lying around.
313 		 */
314 		if (state != 0)
315 			return (-1);
316 	}
317 
318 	return (tarindex);
319 }
320