xref: /illumos-gate/usr/src/lib/libresolv2/common/isc/base64.c (revision e5803b76927480e8f9b67b22201c484ccf4c2bcf)
1 /*
2  * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
3  * Use is subject to license terms.
4  */
5 
6 
7 /*
8  * Copyright (c) 2004 by Internet Systems Consortium, Inc. ("ISC")
9  * Copyright (c) 1996-1999 by Internet Software Consortium.
10  *
11  * Permission to use, copy, modify, and distribute this software for any
12  * purpose with or without fee is hereby granted, provided that the above
13  * copyright notice and this permission notice appear in all copies.
14  *
15  * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES
16  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
17  * MERCHANTABILITY AND FITNESS.  IN NO EVENT SHALL ISC BE LIABLE FOR
18  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
19  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
20  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
21  * OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
22  */
23 
24 /*
25  * Portions Copyright (c) 1995 by International Business Machines, Inc.
26  *
27  * International Business Machines, Inc. (hereinafter called IBM) grants
28  * permission under its copyrights to use, copy, modify, and distribute this
29  * Software with or without fee, provided that the above copyright notice and
30  * all paragraphs of this notice appear in all copies, and that the name of IBM
31  * not be used in connection with the marketing of any product incorporating
32  * the Software or modifications thereof, without specific, written prior
33  * permission.
34  *
35  * To the extent it has a right to do so, IBM grants an immunity from suit
36  * under its patents, if any, for the use, sale or manufacture of products to
37  * the extent that such products are used for performing Domain Name System
38  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
39  * granted for any product per se or for any other function of any product.
40  *
41  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
42  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
43  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
44  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
45  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
46  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
47  */
48 
49 #if !defined(LINT) && !defined(CODECENTER)
50 static const char rcsid[] = "$Id: base64.c,v 1.4 2005/04/27 04:56:34 sra Exp $";
51 #endif /* not lint */
52 
53 #include "port_before.h"
54 
55 #include <sys/types.h>
56 #include <sys/param.h>
57 #include <sys/socket.h>
58 
59 #include <netinet/in.h>
60 #include <arpa/inet.h>
61 #include <arpa/nameser.h>
62 
63 #include <ctype.h>
64 #include <resolv.h>
65 #include <stdio.h>
66 #include <stdlib.h>
67 #include <string.h>
68 
69 #include "port_after.h"
70 
71 #ifndef	ORIGINAL_ISC_CODE
72 #pragma weak	__b64_ntop	=	b64_ntop
73 #pragma weak	__b64_pton	=	b64_pton
74 #endif	/* ORIGINAL_ISC_CODE */
75 
76 #define Assert(Cond) if (!(Cond)) abort()
77 
78 static const char Base64[] =
79 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
80 static const char Pad64 = '=';
81 
82 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
83    The following encoding technique is taken from RFC1521 by Borenstein
84    and Freed.  It is reproduced here in a slightly edited form for
85    convenience.
86 
87    A 65-character subset of US-ASCII is used, enabling 6 bits to be
88    represented per printable character. (The extra 65th character, "=",
89    is used to signify a special processing function.)
90 
91    The encoding process represents 24-bit groups of input bits as output
92    strings of 4 encoded characters. Proceeding from left to right, a
93    24-bit input group is formed by concatenating 3 8-bit input groups.
94    These 24 bits are then treated as 4 concatenated 6-bit groups, each
95    of which is translated into a single digit in the base64 alphabet.
96 
97    Each 6-bit group is used as an index into an array of 64 printable
98    characters. The character referenced by the index is placed in the
99    output string.
100 
101                          Table 1: The Base64 Alphabet
102 
103       Value Encoding  Value Encoding  Value Encoding  Value Encoding
104           0 A            17 R            34 i            51 z
105           1 B            18 S            35 j            52 0
106           2 C            19 T            36 k            53 1
107           3 D            20 U            37 l            54 2
108           4 E            21 V            38 m            55 3
109           5 F            22 W            39 n            56 4
110           6 G            23 X            40 o            57 5
111           7 H            24 Y            41 p            58 6
112           8 I            25 Z            42 q            59 7
113           9 J            26 a            43 r            60 8
114          10 K            27 b            44 s            61 9
115          11 L            28 c            45 t            62 +
116          12 M            29 d            46 u            63 /
117          13 N            30 e            47 v
118          14 O            31 f            48 w         (pad) =
119          15 P            32 g            49 x
120          16 Q            33 h            50 y
121 
122    Special processing is performed if fewer than 24 bits are available
123    at the end of the data being encoded.  A full encoding quantum is
124    always completed at the end of a quantity.  When fewer than 24 input
125    bits are available in an input group, zero bits are added (on the
126    right) to form an integral number of 6-bit groups.  Padding at the
127    end of the data is performed using the '=' character.
128 
129    Since all base64 input is an integral number of octets, only the
130          -------------------------------------------------
131    following cases can arise:
132 
133        (1) the final quantum of encoding input is an integral
134            multiple of 24 bits; here, the final unit of encoded
135 	   output will be an integral multiple of 4 characters
136 	   with no "=" padding,
137        (2) the final quantum of encoding input is exactly 8 bits;
138            here, the final unit of encoded output will be two
139 	   characters followed by two "=" padding characters, or
140        (3) the final quantum of encoding input is exactly 16 bits;
141            here, the final unit of encoded output will be three
142 	   characters followed by one "=" padding character.
143    */
144 
145 int
146 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
147 	size_t datalength = 0;
148 	u_char input[3];
149 	u_char output[4];
150 	size_t i;
151 
152 	while (2U < srclength) {
153 		input[0] = *src++;
154 		input[1] = *src++;
155 		input[2] = *src++;
156 		srclength -= 3;
157 
158 		output[0] = input[0] >> 2;
159 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
160 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
161 		output[3] = input[2] & 0x3f;
162 		Assert(output[0] < 64);
163 		Assert(output[1] < 64);
164 		Assert(output[2] < 64);
165 		Assert(output[3] < 64);
166 
167 		if (datalength + 4 > targsize)
168 			return (-1);
169 		target[datalength++] = Base64[output[0]];
170 		target[datalength++] = Base64[output[1]];
171 		target[datalength++] = Base64[output[2]];
172 		target[datalength++] = Base64[output[3]];
173 	}
174 
175 	/* Now we worry about padding. */
176 	if (0U != srclength) {
177 		/* Get what's left. */
178 		input[0] = input[1] = input[2] = '\0';
179 		for (i = 0; i < srclength; i++)
180 			input[i] = *src++;
181 
182 		output[0] = input[0] >> 2;
183 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
184 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
185 		Assert(output[0] < 64);
186 		Assert(output[1] < 64);
187 		Assert(output[2] < 64);
188 
189 		if (datalength + 4 > targsize)
190 			return (-1);
191 		target[datalength++] = Base64[output[0]];
192 		target[datalength++] = Base64[output[1]];
193 		if (srclength == 1U)
194 			target[datalength++] = Pad64;
195 		else
196 			target[datalength++] = Base64[output[2]];
197 		target[datalength++] = Pad64;
198 	}
199 	if (datalength >= targsize)
200 		return (-1);
201 	target[datalength] = '\0';	/*%< Returned value doesn't count \\0. */
202 	return (datalength);
203 }
204 
205 /* skips all whitespace anywhere.
206    converts characters, four at a time, starting at (or after)
207    src from base - 64 numbers into three 8 bit bytes in the target area.
208    it returns the number of data bytes stored at the target, or -1 on error.
209  */
210 
211 int
212 b64_pton(src, target, targsize)
213 	char const *src;
214 	u_char *target;
215 	size_t targsize;
216 {
217 	int tarindex, state, ch;
218 	char *pos;
219 
220 	state = 0;
221 	tarindex = 0;
222 
223 	while ((ch = *src++) != '\0') {
224 		if (isspace(ch))	/*%< Skip whitespace anywhere. */
225 			continue;
226 
227 		if (ch == Pad64)
228 			break;
229 
230 		pos = strchr(Base64, ch);
231 		if (pos == 0) 		/*%< A non-base64 character. */
232 			return (-1);
233 
234 		switch (state) {
235 		case 0:
236 			if (target) {
237 				if ((size_t)tarindex >= targsize)
238 					return (-1);
239 				target[tarindex] = (pos - Base64) << 2;
240 			}
241 			state = 1;
242 			break;
243 		case 1:
244 			if (target) {
245 				if ((size_t)tarindex + 1 >= targsize)
246 					return (-1);
247 				target[tarindex]   |=  (pos - Base64) >> 4;
248 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
249 							<< 4 ;
250 			}
251 			tarindex++;
252 			state = 2;
253 			break;
254 		case 2:
255 			if (target) {
256 				if ((size_t)tarindex + 1 >= targsize)
257 					return (-1);
258 				target[tarindex]   |=  (pos - Base64) >> 2;
259 				target[tarindex+1]  = ((pos - Base64) & 0x03)
260 							<< 6;
261 			}
262 			tarindex++;
263 			state = 3;
264 			break;
265 		case 3:
266 			if (target) {
267 				if ((size_t)tarindex >= targsize)
268 					return (-1);
269 				target[tarindex] |= (pos - Base64);
270 			}
271 			tarindex++;
272 			state = 0;
273 			break;
274 		default:
275 			abort();
276 		}
277 	}
278 
279 	/*
280 	 * We are done decoding Base-64 chars.  Let's see if we ended
281 	 * on a byte boundary, and/or with erroneous trailing characters.
282 	 */
283 
284 	if (ch == Pad64) {		/*%< We got a pad char. */
285 		ch = *src++;		/*%< Skip it, get next. */
286 		switch (state) {
287 		case 0:		/*%< Invalid = in first position */
288 		case 1:		/*%< Invalid = in second position */
289 			return (-1);
290 
291 		case 2:		/*%< Valid, means one byte of info */
292 			/* Skip any number of spaces. */
293 			for ((void)NULL; ch != '\0'; ch = *src++)
294 				if (!isspace(ch))
295 					break;
296 			/* Make sure there is another trailing = sign. */
297 			if (ch != Pad64)
298 				return (-1);
299 			ch = *src++;		/*%< Skip the = */
300 			/* Fall through to "single trailing =" case. */
301 			/* FALLTHROUGH */
302 
303 		case 3:		/*%< Valid, means two bytes of info */
304 			/*
305 			 * We know this char is an =.  Is there anything but
306 			 * whitespace after it?
307 			 */
308 			for ((void)NULL; ch != '\0'; ch = *src++)
309 				if (!isspace(ch))
310 					return (-1);
311 
312 			/*
313 			 * Now make sure for cases 2 and 3 that the "extra"
314 			 * bits that slopped past the last full byte were
315 			 * zeros.  If we don't check them, they become a
316 			 * subliminal channel.
317 			 */
318 			if (target && target[tarindex] != 0)
319 				return (-1);
320 		}
321 	} else {
322 		/*
323 		 * We ended by seeing the end of the string.  Make sure we
324 		 * have no partial bytes lying around.
325 		 */
326 		if (state != 0)
327 			return (-1);
328 	}
329 
330 	return (tarindex);
331 }
332 
333 /*! \file */
334