1 /*	$OpenBSD: base64.c,v 1.5 2006/10/21 09:55:03 otto Exp $	*/
2 
3 /*
4  * Copyright (c) 1996 by Internet Software Consortium.
5  *
6  * Permission to use, copy, modify, and distribute this software for any
7  * purpose with or without fee is hereby granted, provided that the above
8  * copyright notice and this permission notice appear in all copies.
9  *
10  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
11  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
12  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
13  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
14  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
15  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
16  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
17  * 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 /* OPENBSD ORIGINAL: lib/libc/net/base64.c */
46 
47 #include "includes.h"
48 
49 #if (!defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)) || (!defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON))
50 
51 #include <sys/types.h>
52 #include <sys/param.h>
53 #include <sys/socket.h>
54 #include <netinet/in.h>
55 #include <arpa/inet.h>
56 
57 #include <ctype.h>
58 #include <stdio.h>
59 
60 #include <stdlib.h>
61 #include <string.h>
62 
63 #include "base64.h"
64 
65 static const char Base64[] =
66 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
67 static const char Pad64 = '=';
68 
69 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
70    The following encoding technique is taken from RFC 1521 by Borenstein
71    and Freed.  It is reproduced here in a slightly edited form for
72    convenience.
73 
74    A 65-character subset of US-ASCII is used, enabling 6 bits to be
75    represented per printable character. (The extra 65th character, "=",
76    is used to signify a special processing function.)
77 
78    The encoding process represents 24-bit groups of input bits as output
79    strings of 4 encoded characters. Proceeding from left to right, a
80    24-bit input group is formed by concatenating 3 8-bit input groups.
81    These 24 bits are then treated as 4 concatenated 6-bit groups, each
82    of which is translated into a single digit in the base64 alphabet.
83 
84    Each 6-bit group is used as an index into an array of 64 printable
85    characters. The character referenced by the index is placed in the
86    output string.
87 
88                          Table 1: The Base64 Alphabet
89 
90       Value Encoding  Value Encoding  Value Encoding  Value Encoding
91           0 A            17 R            34 i            51 z
92           1 B            18 S            35 j            52 0
93           2 C            19 T            36 k            53 1
94           3 D            20 U            37 l            54 2
95           4 E            21 V            38 m            55 3
96           5 F            22 W            39 n            56 4
97           6 G            23 X            40 o            57 5
98           7 H            24 Y            41 p            58 6
99           8 I            25 Z            42 q            59 7
100           9 J            26 a            43 r            60 8
101          10 K            27 b            44 s            61 9
102          11 L            28 c            45 t            62 +
103          12 M            29 d            46 u            63 /
104          13 N            30 e            47 v
105          14 O            31 f            48 w         (pad) =
106          15 P            32 g            49 x
107          16 Q            33 h            50 y
108 
109    Special processing is performed if fewer than 24 bits are available
110    at the end of the data being encoded.  A full encoding quantum is
111    always completed at the end of a quantity.  When fewer than 24 input
112    bits are available in an input group, zero bits are added (on the
113    right) to form an integral number of 6-bit groups.  Padding at the
114    end of the data is performed using the '=' character.
115 
116    Since all base64 input is an integral number of octets, only the
117          -------------------------------------------------
118    following cases can arise:
119 
120        (1) the final quantum of encoding input is an integral
121            multiple of 24 bits; here, the final unit of encoded
122 	   output will be an integral multiple of 4 characters
123 	   with no "=" padding,
124        (2) the final quantum of encoding input is exactly 8 bits;
125            here, the final unit of encoded output will be two
126 	   characters followed by two "=" padding characters, or
127        (3) the final quantum of encoding input is exactly 16 bits;
128            here, the final unit of encoded output will be three
129 	   characters followed by one "=" padding character.
130    */
131 
132 #if !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)
133 int
b64_ntop(u_char const * src,size_t srclength,char * target,size_t targsize)134 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize)
135 {
136 	size_t datalength = 0;
137 	u_char input[3];
138 	u_char output[4];
139 	u_int i;
140 
141 	while (2 < srclength) {
142 		input[0] = *src++;
143 		input[1] = *src++;
144 		input[2] = *src++;
145 		srclength -= 3;
146 
147 		output[0] = input[0] >> 2;
148 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
149 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
150 		output[3] = input[2] & 0x3f;
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 
171 		if (datalength + 4 > targsize)
172 			return (-1);
173 		target[datalength++] = Base64[output[0]];
174 		target[datalength++] = Base64[output[1]];
175 		if (srclength == 1)
176 			target[datalength++] = Pad64;
177 		else
178 			target[datalength++] = Base64[output[2]];
179 		target[datalength++] = Pad64;
180 	}
181 	if (datalength >= targsize)
182 		return (-1);
183 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
184 	return (datalength);
185 }
186 #endif /* !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) */
187 
188 #if !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON)
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
b64_pton(char const * src,u_char * target,size_t targsize)197 b64_pton(char const *src, u_char *target, size_t targsize)
198 {
199 	u_int tarindex, state;
200 	int ch;
201 	char *pos;
202 
203 	state = 0;
204 	tarindex = 0;
205 
206 	while ((ch = *src++) != '\0') {
207 		if (isspace(ch))	/* Skip whitespace anywhere. */
208 			continue;
209 
210 		if (ch == Pad64)
211 			break;
212 
213 		pos = strchr(Base64, ch);
214 		if (pos == 0)		/* A non-base64 character. */
215 			return (-1);
216 
217 		switch (state) {
218 		case 0:
219 			if (target) {
220 				if (tarindex >= targsize)
221 					return (-1);
222 				target[tarindex] = (pos - Base64) << 2;
223 			}
224 			state = 1;
225 			break;
226 		case 1:
227 			if (target) {
228 				if (tarindex + 1 >= targsize)
229 					return (-1);
230 				target[tarindex]   |=  (pos - Base64) >> 4;
231 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
232 							<< 4 ;
233 			}
234 			tarindex++;
235 			state = 2;
236 			break;
237 		case 2:
238 			if (target) {
239 				if (tarindex + 1 >= targsize)
240 					return (-1);
241 				target[tarindex]   |=  (pos - Base64) >> 2;
242 				target[tarindex+1]  = ((pos - Base64) & 0x03)
243 							<< 6;
244 			}
245 			tarindex++;
246 			state = 3;
247 			break;
248 		case 3:
249 			if (target) {
250 				if (tarindex >= targsize)
251 					return (-1);
252 				target[tarindex] |= (pos - Base64);
253 			}
254 			tarindex++;
255 			state = 0;
256 			break;
257 		}
258 	}
259 
260 	/*
261 	 * We are done decoding Base-64 chars.  Let's see if we ended
262 	 * on a byte boundary, and/or with erroneous trailing characters.
263 	 */
264 
265 	if (ch == Pad64) {		/* We got a pad char. */
266 		ch = *src++;		/* Skip it, get next. */
267 		switch (state) {
268 		case 0:		/* Invalid = in first position */
269 		case 1:		/* Invalid = in second position */
270 			return (-1);
271 
272 		case 2:		/* Valid, means one byte of info */
273 			/* Skip any number of spaces. */
274 			for (; ch != '\0'; ch = *src++)
275 				if (!isspace(ch))
276 					break;
277 			/* Make sure there is another trailing = sign. */
278 			if (ch != Pad64)
279 				return (-1);
280 			ch = *src++;		/* Skip the = */
281 			/* Fall through to "single trailing =" case. */
282 			/* FALLTHROUGH */
283 
284 		case 3:		/* Valid, means two bytes of info */
285 			/*
286 			 * We know this char is an =.  Is there anything but
287 			 * whitespace after it?
288 			 */
289 			for (; ch != '\0'; ch = *src++)
290 				if (!isspace(ch))
291 					return (-1);
292 
293 			/*
294 			 * Now make sure for cases 2 and 3 that the "extra"
295 			 * bits that slopped past the last full byte were
296 			 * zeros.  If we don't check them, they become a
297 			 * subliminal channel.
298 			 */
299 			if (target && target[tarindex] != 0)
300 				return (-1);
301 		}
302 	} else {
303 		/*
304 		 * We ended by seeing the end of the string.  Make sure we
305 		 * have no partial bytes lying around.
306 		 */
307 		if (state != 0)
308 			return (-1);
309 	}
310 
311 	return (tarindex);
312 }
313 
314 #endif /* !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON) */
315 #endif
316