xref: /netbsd-src/external/bsd/libbind/dist/isc/base64.c (revision 5bbd2a12505d72a8177929a37b5cee489d0a1cfd)
1 /*	$NetBSD: base64.c,v 1.1.1.2 2012/09/09 16:08:02 christos Exp $	*/
2 
3 /*
4  * Copyright (c) 2004 by Internet Systems Consortium, Inc. ("ISC")
5  * Copyright (c) 1996-1999 by Internet Software Consortium.
6  *
7  * Permission to use, copy, modify, and distribute this software for any
8  * purpose with or without fee is hereby granted, provided that the above
9  * copyright notice and this permission notice appear in all copies.
10  *
11  * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES
12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
13  * MERCHANTABILITY AND FITNESS.  IN NO EVENT SHALL ISC BE LIABLE FOR
14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
17  * OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18  */
19 
20 /*
21  * Portions Copyright (c) 1995 by International Business Machines, Inc.
22  *
23  * International Business Machines, Inc. (hereinafter called IBM) grants
24  * permission under its copyrights to use, copy, modify, and distribute this
25  * Software with or without fee, provided that the above copyright notice and
26  * all paragraphs of this notice appear in all copies, and that the name of IBM
27  * not be used in connection with the marketing of any product incorporating
28  * the Software or modifications thereof, without specific, written prior
29  * permission.
30  *
31  * To the extent it has a right to do so, IBM grants an immunity from suit
32  * under its patents, if any, for the use, sale or manufacture of products to
33  * the extent that such products are used for performing Domain Name System
34  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
35  * granted for any product per se or for any other function of any product.
36  *
37  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
38  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
39  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
40  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
41  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
42  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
43  */
44 
45 #if !defined(LINT) && !defined(CODECENTER)
46 static const char rcsid[] = "Id: base64.c,v 1.4 2005/04/27 04:56:34 sra Exp ";
47 #endif /* not lint */
48 
49 #include "port_before.h"
50 
51 #include <sys/types.h>
52 #include <sys/param.h>
53 #include <sys/socket.h>
54 
55 #include <netinet/in.h>
56 #include <arpa/inet.h>
57 #include <arpa/nameser.h>
58 
59 #include <ctype.h>
60 #include <resolv.h>
61 #include <stdio.h>
62 #include <stdlib.h>
63 #include <string.h>
64 
65 #include "port_after.h"
66 
67 #define Assert(Cond) if (!(Cond)) abort()
68 
69 static const char Base64[] =
70 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
71 static const char Pad64 = '=';
72 
73 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
74    The following encoding technique is taken from RFC1521 by Borenstein
75    and Freed.  It is reproduced here in a slightly edited form for
76    convenience.
77 
78    A 65-character subset of US-ASCII is used, enabling 6 bits to be
79    represented per printable character. (The extra 65th character, "=",
80    is used to signify a special processing function.)
81 
82    The encoding process represents 24-bit groups of input bits as output
83    strings of 4 encoded characters. Proceeding from left to right, a
84    24-bit input group is formed by concatenating 3 8-bit input groups.
85    These 24 bits are then treated as 4 concatenated 6-bit groups, each
86    of which is translated into a single digit in the base64 alphabet.
87 
88    Each 6-bit group is used as an index into an array of 64 printable
89    characters. The character referenced by the index is placed in the
90    output string.
91 
92                          Table 1: The Base64 Alphabet
93 
94       Value Encoding  Value Encoding  Value Encoding  Value Encoding
95           0 A            17 R            34 i            51 z
96           1 B            18 S            35 j            52 0
97           2 C            19 T            36 k            53 1
98           3 D            20 U            37 l            54 2
99           4 E            21 V            38 m            55 3
100           5 F            22 W            39 n            56 4
101           6 G            23 X            40 o            57 5
102           7 H            24 Y            41 p            58 6
103           8 I            25 Z            42 q            59 7
104           9 J            26 a            43 r            60 8
105          10 K            27 b            44 s            61 9
106          11 L            28 c            45 t            62 +
107          12 M            29 d            46 u            63 /
108          13 N            30 e            47 v
109          14 O            31 f            48 w         (pad) =
110          15 P            32 g            49 x
111          16 Q            33 h            50 y
112 
113    Special processing is performed if fewer than 24 bits are available
114    at the end of the data being encoded.  A full encoding quantum is
115    always completed at the end of a quantity.  When fewer than 24 input
116    bits are available in an input group, zero bits are added (on the
117    right) to form an integral number of 6-bit groups.  Padding at the
118    end of the data is performed using the '=' character.
119 
120    Since all base64 input is an integral number of octets, only the
121          -------------------------------------------------
122    following cases can arise:
123 
124        (1) the final quantum of encoding input is an integral
125            multiple of 24 bits; here, the final unit of encoded
126 	   output will be an integral multiple of 4 characters
127 	   with no "=" padding,
128        (2) the final quantum of encoding input is exactly 8 bits;
129            here, the final unit of encoded output will be two
130 	   characters followed by two "=" padding characters, or
131        (3) the final quantum of encoding input is exactly 16 bits;
132            here, the final unit of encoded output will be three
133 	   characters followed by one "=" padding character.
134    */
135 
136 int
b64_ntop(u_char const * src,size_t srclength,char * target,size_t targsize)137 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
138 	size_t datalength = 0;
139 	u_char input[3];
140 	u_char output[4];
141 	size_t i;
142 
143 	while (2U < srclength) {
144 		input[0] = *src++;
145 		input[1] = *src++;
146 		input[2] = *src++;
147 		srclength -= 3;
148 
149 		output[0] = input[0] >> 2;
150 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
151 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
152 		output[3] = input[2] & 0x3f;
153 		Assert(output[0] < 64);
154 		Assert(output[1] < 64);
155 		Assert(output[2] < 64);
156 		Assert(output[3] < 64);
157 
158 		if (datalength + 4 > targsize)
159 			return (-1);
160 		target[datalength++] = Base64[output[0]];
161 		target[datalength++] = Base64[output[1]];
162 		target[datalength++] = Base64[output[2]];
163 		target[datalength++] = Base64[output[3]];
164 	}
165 
166 	/* Now we worry about padding. */
167 	if (0U != srclength) {
168 		/* Get what's left. */
169 		input[0] = input[1] = input[2] = '\0';
170 		for (i = 0; i < srclength; i++)
171 			input[i] = *src++;
172 
173 		output[0] = input[0] >> 2;
174 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
175 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
176 		Assert(output[0] < 64);
177 		Assert(output[1] < 64);
178 		Assert(output[2] < 64);
179 
180 		if (datalength + 4 > targsize)
181 			return (-1);
182 		target[datalength++] = Base64[output[0]];
183 		target[datalength++] = Base64[output[1]];
184 		if (srclength == 1U)
185 			target[datalength++] = Pad64;
186 		else
187 			target[datalength++] = Base64[output[2]];
188 		target[datalength++] = Pad64;
189 	}
190 	if (datalength >= targsize)
191 		return (-1);
192 	target[datalength] = '\0';	/*%< Returned value doesn't count \\0. */
193 	return (datalength);
194 }
195 
196 /* skips all whitespace anywhere.
197    converts characters, four at a time, starting at (or after)
198    src from base - 64 numbers into three 8 bit bytes in the target area.
199    it returns the number of data bytes stored at the target, or -1 on error.
200  */
201 
202 int
b64_pton(src,target,targsize)203 b64_pton(src, target, targsize)
204 	char const *src;
205 	u_char *target;
206 	size_t targsize;
207 {
208 	int tarindex, state, ch;
209 	char *pos;
210 
211 	state = 0;
212 	tarindex = 0;
213 
214 	while ((ch = *src++) != '\0') {
215 		if (isspace(ch))	/*%< Skip whitespace anywhere. */
216 			continue;
217 
218 		if (ch == Pad64)
219 			break;
220 
221 		pos = strchr(Base64, ch);
222 		if (pos == 0) 		/*%< A non-base64 character. */
223 			return (-1);
224 
225 		switch (state) {
226 		case 0:
227 			if (target) {
228 				if ((size_t)tarindex >= targsize)
229 					return (-1);
230 				target[tarindex] = (pos - Base64) << 2;
231 			}
232 			state = 1;
233 			break;
234 		case 1:
235 			if (target) {
236 				if ((size_t)tarindex + 1 >= targsize)
237 					return (-1);
238 				target[tarindex]   |=  (pos - Base64) >> 4;
239 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
240 							<< 4 ;
241 			}
242 			tarindex++;
243 			state = 2;
244 			break;
245 		case 2:
246 			if (target) {
247 				if ((size_t)tarindex + 1 >= targsize)
248 					return (-1);
249 				target[tarindex]   |=  (pos - Base64) >> 2;
250 				target[tarindex+1]  = ((pos - Base64) & 0x03)
251 							<< 6;
252 			}
253 			tarindex++;
254 			state = 3;
255 			break;
256 		case 3:
257 			if (target) {
258 				if ((size_t)tarindex >= targsize)
259 					return (-1);
260 				target[tarindex] |= (pos - Base64);
261 			}
262 			tarindex++;
263 			state = 0;
264 			break;
265 		default:
266 			abort();
267 		}
268 	}
269 
270 	/*
271 	 * We are done decoding Base-64 chars.  Let's see if we ended
272 	 * on a byte boundary, and/or with erroneous trailing characters.
273 	 */
274 
275 	if (ch == Pad64) {		/*%< We got a pad char. */
276 		ch = *src++;		/*%< Skip it, get next. */
277 		switch (state) {
278 		case 0:		/*%< Invalid = in first position */
279 		case 1:		/*%< Invalid = in second position */
280 			return (-1);
281 
282 		case 2:		/*%< Valid, means one byte of info */
283 			/* Skip any number of spaces. */
284 			for ((void)NULL; ch != '\0'; ch = *src++)
285 				if (!isspace(ch))
286 					break;
287 			/* Make sure there is another trailing = sign. */
288 			if (ch != Pad64)
289 				return (-1);
290 			ch = *src++;		/*%< Skip the = */
291 			/* Fall through to "single trailing =" case. */
292 			/* FALLTHROUGH */
293 
294 		case 3:		/*%< Valid, means two bytes of info */
295 			/*
296 			 * We know this char is an =.  Is there anything but
297 			 * whitespace after it?
298 			 */
299 			for ((void)NULL; ch != '\0'; ch = *src++)
300 				if (!isspace(ch))
301 					return (-1);
302 
303 			/*
304 			 * Now make sure for cases 2 and 3 that the "extra"
305 			 * bits that slopped past the last full byte were
306 			 * zeros.  If we don't check them, they become a
307 			 * subliminal channel.
308 			 */
309 			if (target && target[tarindex] != 0)
310 				return (-1);
311 		}
312 	} else {
313 		/*
314 		 * We ended by seeing the end of the string.  Make sure we
315 		 * have no partial bytes lying around.
316 		 */
317 		if (state != 0)
318 			return (-1);
319 	}
320 
321 	return (tarindex);
322 }
323 
324 /*! \file */
325