xref: /dragonfly/lib/libc/net/base64.c (revision e3146d3a)
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 /*
44  * $FreeBSD: src/lib/libc/net/base64.c,v 1.4 1999/11/04 04:30:43 ache Exp $
45  * $DragonFly: src/lib/libc/net/base64.c,v 1.3 2005/09/19 09:25:15 asmodai Exp $
46  */
47 
48 #include <sys/types.h>
49 #include <sys/param.h>
50 #include <sys/socket.h>
51 
52 #include <netinet/in.h>
53 #include <arpa/inet.h>
54 #include <arpa/nameser.h>
55 
56 #include <ctype.h>
57 #include <resolv.h>
58 #include <stdio.h>
59 #include <stdlib.h>
60 #include <string.h>
61 
62 #define Assert(Cond) if (!(Cond)) abort()
63 
64 static const char Base64[] =
65 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
66 static const char Pad64 = '=';
67 
68 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
69    The following encoding technique is taken from RFC 1521 by Borenstein
70    and Freed.  It is reproduced here in a slightly edited form for
71    convenience.
72 
73    A 65-character subset of US-ASCII is used, enabling 6 bits to be
74    represented per printable character. (The extra 65th character, "=",
75    is used to signify a special processing function.)
76 
77    The encoding process represents 24-bit groups of input bits as output
78    strings of 4 encoded characters. Proceeding from left to right, a
79    24-bit input group is formed by concatenating 3 8-bit input groups.
80    These 24 bits are then treated as 4 concatenated 6-bit groups, each
81    of which is translated into a single digit in the base64 alphabet.
82 
83    Each 6-bit group is used as an index into an array of 64 printable
84    characters. The character referenced by the index is placed in the
85    output string.
86 
87                          Table 1: The Base64 Alphabet
88 
89       Value Encoding  Value Encoding  Value Encoding  Value Encoding
90           0 A            17 R            34 i            51 z
91           1 B            18 S            35 j            52 0
92           2 C            19 T            36 k            53 1
93           3 D            20 U            37 l            54 2
94           4 E            21 V            38 m            55 3
95           5 F            22 W            39 n            56 4
96           6 G            23 X            40 o            57 5
97           7 H            24 Y            41 p            58 6
98           8 I            25 Z            42 q            59 7
99           9 J            26 a            43 r            60 8
100          10 K            27 b            44 s            61 9
101          11 L            28 c            45 t            62 +
102          12 M            29 d            46 u            63 /
103          13 N            30 e            47 v
104          14 O            31 f            48 w         (pad) =
105          15 P            32 g            49 x
106          16 Q            33 h            50 y
107 
108    Special processing is performed if fewer than 24 bits are available
109    at the end of the data being encoded.  A full encoding quantum is
110    always completed at the end of a quantity.  When fewer than 24 input
111    bits are available in an input group, zero bits are added (on the
112    right) to form an integral number of 6-bit groups.  Padding at the
113    end of the data is performed using the '=' character.
114 
115    Since all base64 input is an integral number of octets, only the
116          -------------------------------------------------
117    following cases can arise:
118 
119        (1) the final quantum of encoding input is an integral
120            multiple of 24 bits; here, the final unit of encoded
121 	   output will be an integral multiple of 4 characters
122 	   with no "=" padding,
123        (2) the final quantum of encoding input is exactly 8 bits;
124            here, the final unit of encoded output will be two
125 	   characters followed by two "=" padding characters, or
126        (3) the final quantum of encoding input is exactly 16 bits;
127            here, the final unit of encoded output will be three
128 	   characters followed by one "=" padding character.
129    */
130 
131 int
132 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize) {
133 	size_t datalength = 0;
134 	u_char input[3];
135 	u_char output[4];
136 	size_t i;
137 
138 	while (2 < srclength) {
139 		input[0] = *src++;
140 		input[1] = *src++;
141 		input[2] = *src++;
142 		srclength -= 3;
143 
144 		output[0] = input[0] >> 2;
145 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
146 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
147 		output[3] = input[2] & 0x3f;
148 		Assert(output[0] < 64);
149 		Assert(output[1] < 64);
150 		Assert(output[2] < 64);
151 		Assert(output[3] < 64);
152 
153 		if (datalength + 4 > targsize)
154 			return (-1);
155 		target[datalength++] = Base64[output[0]];
156 		target[datalength++] = Base64[output[1]];
157 		target[datalength++] = Base64[output[2]];
158 		target[datalength++] = Base64[output[3]];
159 	}
160 
161 	/* Now we worry about padding. */
162 	if (0 != srclength) {
163 		/* Get what's left. */
164 		input[0] = input[1] = input[2] = '\0';
165 		for (i = 0; i < srclength; i++)
166 			input[i] = *src++;
167 
168 		output[0] = input[0] >> 2;
169 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
170 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
171 		Assert(output[0] < 64);
172 		Assert(output[1] < 64);
173 		Assert(output[2] < 64);
174 
175 		if (datalength + 4 > targsize)
176 			return (-1);
177 		target[datalength++] = Base64[output[0]];
178 		target[datalength++] = Base64[output[1]];
179 		if (srclength == 1)
180 			target[datalength++] = Pad64;
181 		else
182 			target[datalength++] = Base64[output[2]];
183 		target[datalength++] = Pad64;
184 	}
185 	if (datalength >= targsize)
186 		return (-1);
187 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
188 	return (datalength);
189 }
190 
191 /* skips all whitespace anywhere.
192    converts characters, four at a time, starting at (or after)
193    src from base - 64 numbers into three 8 bit bytes in the target area.
194    it returns the number of data bytes stored at the target, or -1 on error.
195  */
196 
197 int
198 b64_pton(src, target, targsize)
199 	char const *src;
200 	u_char *target;
201 	size_t targsize;
202 {
203 	int tarindex, state, ch;
204 	char *pos;
205 
206 	state = 0;
207 	tarindex = 0;
208 
209 	while ((ch = *src++) != '\0') {
210 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
211 			continue;
212 
213 		if (ch == Pad64)
214 			break;
215 
216 		pos = strchr(Base64, ch);
217 		if (pos == 0) 		/* A non-base64 character. */
218 			return (-1);
219 
220 		switch (state) {
221 		case 0:
222 			if (target) {
223 				if ((size_t)tarindex >= targsize)
224 					return (-1);
225 				target[tarindex] = (pos - Base64) << 2;
226 			}
227 			state = 1;
228 			break;
229 		case 1:
230 			if (target) {
231 				if ((size_t)tarindex + 1 >= targsize)
232 					return (-1);
233 				target[tarindex]   |=  (pos - Base64) >> 4;
234 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
235 							<< 4 ;
236 			}
237 			tarindex++;
238 			state = 2;
239 			break;
240 		case 2:
241 			if (target) {
242 				if ((size_t)tarindex + 1 >= targsize)
243 					return (-1);
244 				target[tarindex]   |=  (pos - Base64) >> 2;
245 				target[tarindex+1]  = ((pos - Base64) & 0x03)
246 							<< 6;
247 			}
248 			tarindex++;
249 			state = 3;
250 			break;
251 		case 3:
252 			if (target) {
253 				if ((size_t)tarindex >= targsize)
254 					return (-1);
255 				target[tarindex] |= (pos - Base64);
256 			}
257 			tarindex++;
258 			state = 0;
259 			break;
260 		default:
261 			abort();
262 		}
263 	}
264 
265 	/*
266 	 * We are done decoding Base-64 chars.  Let's see if we ended
267 	 * on a byte boundary, and/or with erroneous trailing characters.
268 	 */
269 
270 	if (ch == Pad64) {		/* We got a pad char. */
271 		ch = *src++;		/* Skip it, get next. */
272 		switch (state) {
273 		case 0:		/* Invalid = in first position */
274 		case 1:		/* Invalid = in second position */
275 			return (-1);
276 
277 		case 2:		/* Valid, means one byte of info */
278 			/* Skip any number of spaces. */
279 			for ((void)NULL; ch != '\0'; ch = *src++)
280 				if (!isspace((unsigned char)ch))
281 					break;
282 			/* Make sure there is another trailing = sign. */
283 			if (ch != Pad64)
284 				return (-1);
285 			ch = *src++;		/* Skip the = */
286 			/* Fall through to "single trailing =" case. */
287 			/* FALLTHROUGH */
288 
289 		case 3:		/* Valid, means two bytes of info */
290 			/*
291 			 * We know this char is an =.  Is there anything but
292 			 * whitespace after it?
293 			 */
294 			for ((void)NULL; ch != '\0'; ch = *src++)
295 				if (!isspace((unsigned char)ch))
296 					return (-1);
297 
298 			/*
299 			 * Now make sure for cases 2 and 3 that the "extra"
300 			 * bits that slopped past the last full byte were
301 			 * zeros.  If we don't check them, they become a
302 			 * subliminal channel.
303 			 */
304 			if (target && target[tarindex] != 0)
305 				return (-1);
306 		}
307 	} else {
308 		/*
309 		 * We ended by seeing the end of the string.  Make sure we
310 		 * have no partial bytes lying around.
311 		 */
312 		if (state != 0)
313 			return (-1);
314 	}
315 
316 	return (tarindex);
317 }
318