xref: /dragonfly/lib/libc/net/base64.c (revision b40e316c)
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  * $FreeBSD: src/lib/libc/net/base64.c,v 1.4 1999/11/04 04:30:43 ache Exp $
44  * $DragonFly: src/lib/libc/net/base64.c,v 1.2 2003/06/17 04:26:44 dillon Exp $
45  */
46 
47 #include <sys/types.h>
48 #include <sys/param.h>
49 #include <sys/socket.h>
50 
51 #include <netinet/in.h>
52 #include <arpa/inet.h>
53 #include <arpa/nameser.h>
54 
55 #include <ctype.h>
56 #include <resolv.h>
57 #include <stdio.h>
58 #include <stdlib.h>
59 #include <string.h>
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(u_char const *src, size_t srclength, char *target, size_t targsize) {
132 	size_t datalength = 0;
133 	u_char input[3];
134 	u_char output[4];
135 	size_t i;
136 
137 	while (2 < srclength) {
138 		input[0] = *src++;
139 		input[1] = *src++;
140 		input[2] = *src++;
141 		srclength -= 3;
142 
143 		output[0] = input[0] >> 2;
144 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
145 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
146 		output[3] = input[2] & 0x3f;
147 		Assert(output[0] < 64);
148 		Assert(output[1] < 64);
149 		Assert(output[2] < 64);
150 		Assert(output[3] < 64);
151 
152 		if (datalength + 4 > targsize)
153 			return (-1);
154 		target[datalength++] = Base64[output[0]];
155 		target[datalength++] = Base64[output[1]];
156 		target[datalength++] = Base64[output[2]];
157 		target[datalength++] = Base64[output[3]];
158 	}
159 
160 	/* Now we worry about padding. */
161 	if (0 != srclength) {
162 		/* Get what's left. */
163 		input[0] = input[1] = input[2] = '\0';
164 		for (i = 0; i < srclength; i++)
165 			input[i] = *src++;
166 
167 		output[0] = input[0] >> 2;
168 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
169 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
170 		Assert(output[0] < 64);
171 		Assert(output[1] < 64);
172 		Assert(output[2] < 64);
173 
174 		if (datalength + 4 > targsize)
175 			return (-1);
176 		target[datalength++] = Base64[output[0]];
177 		target[datalength++] = Base64[output[1]];
178 		if (srclength == 1)
179 			target[datalength++] = Pad64;
180 		else
181 			target[datalength++] = Base64[output[2]];
182 		target[datalength++] = Pad64;
183 	}
184 	if (datalength >= targsize)
185 		return (-1);
186 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
187 	return (datalength);
188 }
189 
190 /* skips all whitespace anywhere.
191    converts characters, four at a time, starting at (or after)
192    src from base - 64 numbers into three 8 bit bytes in the target area.
193    it returns the number of data bytes stored at the target, or -1 on error.
194  */
195 
196 int
197 b64_pton(src, target, targsize)
198 	char const *src;
199 	u_char *target;
200 	size_t targsize;
201 {
202 	int tarindex, state, ch;
203 	char *pos;
204 
205 	state = 0;
206 	tarindex = 0;
207 
208 	while ((ch = *src++) != '\0') {
209 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
210 			continue;
211 
212 		if (ch == Pad64)
213 			break;
214 
215 		pos = strchr(Base64, ch);
216 		if (pos == 0) 		/* A non-base64 character. */
217 			return (-1);
218 
219 		switch (state) {
220 		case 0:
221 			if (target) {
222 				if ((size_t)tarindex >= targsize)
223 					return (-1);
224 				target[tarindex] = (pos - Base64) << 2;
225 			}
226 			state = 1;
227 			break;
228 		case 1:
229 			if (target) {
230 				if ((size_t)tarindex + 1 >= targsize)
231 					return (-1);
232 				target[tarindex]   |=  (pos - Base64) >> 4;
233 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
234 							<< 4 ;
235 			}
236 			tarindex++;
237 			state = 2;
238 			break;
239 		case 2:
240 			if (target) {
241 				if ((size_t)tarindex + 1 >= targsize)
242 					return (-1);
243 				target[tarindex]   |=  (pos - Base64) >> 2;
244 				target[tarindex+1]  = ((pos - Base64) & 0x03)
245 							<< 6;
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 && 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