xref: /illumos-gate/usr/src/lib/libresolv2/common/isc/base64.c (revision 581cede61ac9c14d8d4ea452562a567189eead78)
1 /*
2  * Copyright (c) 1997-2000 by Sun Microsystems, Inc.
3  * All rights reserved.
4  */
5 
6 /*
7  * Copyright (c) 1996-1999 by Internet Software Consortium.
8  *
9  * Permission to use, copy, modify, and distribute this software for any
10  * purpose with or without fee is hereby granted, provided that the above
11  * copyright notice and this permission notice appear in all copies.
12  *
13  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
14  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
15  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
16  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
17  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
18  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
19  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
20  * SOFTWARE.
21  */
22 
23 #pragma ident	"%Z%%M%	%I%	%E% SMI"
24 
25 /*
26  * Portions Copyright (c) 1995 by International Business Machines, Inc.
27  *
28  * International Business Machines, Inc. (hereinafter called IBM) grants
29  * permission under its copyrights to use, copy, modify, and distribute this
30  * Software with or without fee, provided that the above copyright notice and
31  * all paragraphs of this notice appear in all copies, and that the name of IBM
32  * not be used in connection with the marketing of any product incorporating
33  * the Software or modifications thereof, without specific, written prior
34  * permission.
35  *
36  * To the extent it has a right to do so, IBM grants an immunity from suit
37  * under its patents, if any, for the use, sale or manufacture of products to
38  * the extent that such products are used for performing Domain Name System
39  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
40  * granted for any product per se or for any other function of any product.
41  *
42  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
43  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
44  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
45  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
46  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
47  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
48  */
49 
50 #if !defined(LINT) && !defined(CODECENTER)
51 static const char rcsid[] = "$Id: base64.c,v 8.7 1999/10/13 16:39:33 vixie Exp $";
52 #endif /* not lint */
53 
54 #include "port_before.h"
55 
56 #include <sys/types.h>
57 #include <sys/param.h>
58 #include <sys/socket.h>
59 
60 #include <netinet/in.h>
61 #include <arpa/inet.h>
62 #include <arpa/nameser.h>
63 
64 #include <ctype.h>
65 #include <resolv.h>
66 #include <stdio.h>
67 #include <stdlib.h>
68 #include <string.h>
69 
70 #include "port_after.h"
71 
72 #ifdef	ORIGINAL_ISC_CODE
73 #else
74 #pragma weak	__b64_ntop	=	b64_ntop
75 #pragma weak	__b64_pton	=	b64_pton
76 #endif	/* ORIGINAL_ISC_CODE */
77 
78 #define Assert(Cond) if (!(Cond)) abort()
79 
80 static const char Base64[] =
81 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
82 static const char Pad64 = '=';
83 
84 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
85    The following encoding technique is taken from RFC 1521 by Borenstein
86    and Freed.  It is reproduced here in a slightly edited form for
87    convenience.
88 
89    A 65-character subset of US-ASCII is used, enabling 6 bits to be
90    represented per printable character. (The extra 65th character, "=",
91    is used to signify a special processing function.)
92 
93    The encoding process represents 24-bit groups of input bits as output
94    strings of 4 encoded characters. Proceeding from left to right, a
95    24-bit input group is formed by concatenating 3 8-bit input groups.
96    These 24 bits are then treated as 4 concatenated 6-bit groups, each
97    of which is translated into a single digit in the base64 alphabet.
98 
99    Each 6-bit group is used as an index into an array of 64 printable
100    characters. The character referenced by the index is placed in the
101    output string.
102 
103                          Table 1: The Base64 Alphabet
104 
105       Value Encoding  Value Encoding  Value Encoding  Value Encoding
106           0 A            17 R            34 i            51 z
107           1 B            18 S            35 j            52 0
108           2 C            19 T            36 k            53 1
109           3 D            20 U            37 l            54 2
110           4 E            21 V            38 m            55 3
111           5 F            22 W            39 n            56 4
112           6 G            23 X            40 o            57 5
113           7 H            24 Y            41 p            58 6
114           8 I            25 Z            42 q            59 7
115           9 J            26 a            43 r            60 8
116          10 K            27 b            44 s            61 9
117          11 L            28 c            45 t            62 +
118          12 M            29 d            46 u            63 /
119          13 N            30 e            47 v
120          14 O            31 f            48 w         (pad) =
121          15 P            32 g            49 x
122          16 Q            33 h            50 y
123 
124    Special processing is performed if fewer than 24 bits are available
125    at the end of the data being encoded.  A full encoding quantum is
126    always completed at the end of a quantity.  When fewer than 24 input
127    bits are available in an input group, zero bits are added (on the
128    right) to form an integral number of 6-bit groups.  Padding at the
129    end of the data is performed using the '=' character.
130 
131    Since all base64 input is an integral number of octets, only the
132          -------------------------------------------------
133    following cases can arise:
134 
135        (1) the final quantum of encoding input is an integral
136            multiple of 24 bits; here, the final unit of encoded
137 	   output will be an integral multiple of 4 characters
138 	   with no "=" padding,
139        (2) the final quantum of encoding input is exactly 8 bits;
140            here, the final unit of encoded output will be two
141 	   characters followed by two "=" padding characters, or
142        (3) the final quantum of encoding input is exactly 16 bits;
143            here, the final unit of encoded output will be three
144 	   characters followed by one "=" padding character.
145    */
146 
147 int
148 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
149 	size_t datalength = 0;
150 	u_char input[3];
151 	u_char output[4];
152 	size_t i;
153 
154 	while (2 < srclength) {
155 		input[0] = *src++;
156 		input[1] = *src++;
157 		input[2] = *src++;
158 		srclength -= 3;
159 
160 		output[0] = input[0] >> 2;
161 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
162 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
163 		output[3] = input[2] & 0x3f;
164 		Assert(output[0] < 64);
165 		Assert(output[1] < 64);
166 		Assert(output[2] < 64);
167 		Assert(output[3] < 64);
168 
169 		if (datalength + 4 > targsize)
170 			return (-1);
171 		target[datalength++] = Base64[output[0]];
172 		target[datalength++] = Base64[output[1]];
173 		target[datalength++] = Base64[output[2]];
174 		target[datalength++] = Base64[output[3]];
175 	}
176 
177 	/* Now we worry about padding. */
178 	if (0 != srclength) {
179 		/* Get what's left. */
180 		input[0] = input[1] = input[2] = '\0';
181 		for (i = 0; i < srclength; i++)
182 			input[i] = *src++;
183 
184 		output[0] = input[0] >> 2;
185 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
186 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
187 		Assert(output[0] < 64);
188 		Assert(output[1] < 64);
189 		Assert(output[2] < 64);
190 
191 		if (datalength + 4 > targsize)
192 			return (-1);
193 		target[datalength++] = Base64[output[0]];
194 		target[datalength++] = Base64[output[1]];
195 		if (srclength == 1)
196 			target[datalength++] = Pad64;
197 		else
198 			target[datalength++] = Base64[output[2]];
199 		target[datalength++] = Pad64;
200 	}
201 	if (datalength >= targsize)
202 		return (-1);
203 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
204 	return (datalength);
205 }
206 
207 /* skips all whitespace anywhere.
208    converts characters, four at a time, starting at (or after)
209    src from base - 64 numbers into three 8 bit bytes in the target area.
210    it returns the number of data bytes stored at the target, or -1 on error.
211  */
212 
213 int
214 b64_pton(src, target, targsize)
215 	char const *src;
216 	u_char *target;
217 	size_t targsize;
218 {
219 	int tarindex, state, ch;
220 	char *pos;
221 
222 	state = 0;
223 	tarindex = 0;
224 
225 	while ((ch = *src++) != '\0') {
226 		if (isspace(ch))	/* Skip whitespace anywhere. */
227 			continue;
228 
229 		if (ch == Pad64)
230 			break;
231 
232 		pos = strchr(Base64, ch);
233 		if (pos == 0) 		/* A non-base64 character. */
234 			return (-1);
235 
236 		switch (state) {
237 		case 0:
238 			if (target) {
239 				if ((size_t)tarindex >= targsize)
240 					return (-1);
241 				target[tarindex] = (pos - Base64) << 2;
242 			}
243 			state = 1;
244 			break;
245 		case 1:
246 			if (target) {
247 				if ((size_t)tarindex + 1 >= targsize)
248 					return (-1);
249 				target[tarindex]   |=  (pos - Base64) >> 4;
250 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
251 							<< 4 ;
252 			}
253 			tarindex++;
254 			state = 2;
255 			break;
256 		case 2:
257 			if (target) {
258 				if ((size_t)tarindex + 1 >= targsize)
259 					return (-1);
260 				target[tarindex]   |=  (pos - Base64) >> 2;
261 				target[tarindex+1]  = ((pos - Base64) & 0x03)
262 							<< 6;
263 			}
264 			tarindex++;
265 			state = 3;
266 			break;
267 		case 3:
268 			if (target) {
269 				if ((size_t)tarindex >= targsize)
270 					return (-1);
271 				target[tarindex] |= (pos - Base64);
272 			}
273 			tarindex++;
274 			state = 0;
275 			break;
276 		default:
277 			abort();
278 		}
279 	}
280 
281 	/*
282 	 * We are done decoding Base-64 chars.  Let's see if we ended
283 	 * on a byte boundary, and/or with erroneous trailing characters.
284 	 */
285 
286 	if (ch == Pad64) {		/* We got a pad char. */
287 		ch = *src++;		/* Skip it, get next. */
288 		switch (state) {
289 		case 0:		/* Invalid = in first position */
290 		case 1:		/* Invalid = in second position */
291 			return (-1);
292 
293 		case 2:		/* Valid, means one byte of info */
294 			/* Skip any number of spaces. */
295 			for ((void)NULL; ch != '\0'; ch = *src++)
296 				if (!isspace(ch))
297 					break;
298 			/* Make sure there is another trailing = sign. */
299 			if (ch != Pad64)
300 				return (-1);
301 			ch = *src++;		/* Skip the = */
302 			/* Fall through to "single trailing =" case. */
303 			/* FALLTHROUGH */
304 
305 		case 3:		/* Valid, means two bytes of info */
306 			/*
307 			 * We know this char is an =.  Is there anything but
308 			 * whitespace after it?
309 			 */
310 			for ((void)NULL; ch != '\0'; ch = *src++)
311 				if (!isspace(ch))
312 					return (-1);
313 
314 			/*
315 			 * Now make sure for cases 2 and 3 that the "extra"
316 			 * bits that slopped past the last full byte were
317 			 * zeros.  If we don't check them, they become a
318 			 * subliminal channel.
319 			 */
320 			if (target && target[tarindex] != 0)
321 				return (-1);
322 		}
323 	} else {
324 		/*
325 		 * We ended by seeing the end of the string.  Make sure we
326 		 * have no partial bytes lying around.
327 		 */
328 		if (state != 0)
329 			return (-1);
330 	}
331 
332 	return (tarindex);
333 }
334