xref: /dragonfly/lib/libc/net/base64.c (revision 335b9e93)
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  */
46 
47 #include <sys/types.h>
48 #include <sys/param.h>
49 #include <sys/socket.h>
50 
51 #include <netinet/in.h>
52 #include <arpa/inet.h>
53 #include <arpa/nameser.h>
54 
55 #include <ctype.h>
56 #include <resolv.h>
57 #include <stdio.h>
58 #include <stdlib.h>
59 #include <string.h>
60 
61 #define Assert(Cond) if (!(Cond)) abort()
62 
63 static const char Base64[] =
64 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
65 static const char Pad64 = '=';
66 
67 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
68    The following encoding technique is taken from RFC 1521 by Borenstein
69    and Freed.  It is reproduced here in a slightly edited form for
70    convenience.
71 
72    A 65-character subset of US-ASCII is used, enabling 6 bits to be
73    represented per printable character. (The extra 65th character, "=",
74    is used to signify a special processing function.)
75 
76    The encoding process represents 24-bit groups of input bits as output
77    strings of 4 encoded characters. Proceeding from left to right, a
78    24-bit input group is formed by concatenating 3 8-bit input groups.
79    These 24 bits are then treated as 4 concatenated 6-bit groups, each
80    of which is translated into a single digit in the base64 alphabet.
81 
82    Each 6-bit group is used as an index into an array of 64 printable
83    characters. The character referenced by the index is placed in the
84    output string.
85 
86                          Table 1: The Base64 Alphabet
87 
88       Value Encoding  Value Encoding  Value Encoding  Value Encoding
89           0 A            17 R            34 i            51 z
90           1 B            18 S            35 j            52 0
91           2 C            19 T            36 k            53 1
92           3 D            20 U            37 l            54 2
93           4 E            21 V            38 m            55 3
94           5 F            22 W            39 n            56 4
95           6 G            23 X            40 o            57 5
96           7 H            24 Y            41 p            58 6
97           8 I            25 Z            42 q            59 7
98           9 J            26 a            43 r            60 8
99          10 K            27 b            44 s            61 9
100          11 L            28 c            45 t            62 +
101          12 M            29 d            46 u            63 /
102          13 N            30 e            47 v
103          14 O            31 f            48 w         (pad) =
104          15 P            32 g            49 x
105          16 Q            33 h            50 y
106 
107    Special processing is performed if fewer than 24 bits are available
108    at the end of the data being encoded.  A full encoding quantum is
109    always completed at the end of a quantity.  When fewer than 24 input
110    bits are available in an input group, zero bits are added (on the
111    right) to form an integral number of 6-bit groups.  Padding at the
112    end of the data is performed using the '=' character.
113 
114    Since all base64 input is an integral number of octets, only the
115          -------------------------------------------------
116    following cases can arise:
117 
118        (1) the final quantum of encoding input is an integral
119            multiple of 24 bits; here, the final unit of encoded
120 	   output will be an integral multiple of 4 characters
121 	   with no "=" padding,
122        (2) the final quantum of encoding input is exactly 8 bits;
123            here, the final unit of encoded output will be two
124 	   characters followed by two "=" padding characters, or
125        (3) the final quantum of encoding input is exactly 16 bits;
126            here, the final unit of encoded output will be three
127 	   characters followed by one "=" padding character.
128    */
129 
130 int
131 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize)
132 {
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(char const *src, u_char *target, size_t targsize)
199 {
200 	int tarindex, state, ch;
201 	char *pos;
202 
203 	state = 0;
204 	tarindex = 0;
205 
206 	while ((ch = *src++) != '\0') {
207 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
208 			continue;
209 
210 		if (ch == Pad64)
211 			break;
212 
213 		pos = strchr(Base64, ch);
214 		if (pos == NULL) 		/* A non-base64 character. */
215 			return (-1);
216 
217 		switch (state) {
218 		case 0:
219 			if (target) {
220 				if ((size_t)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 ((size_t)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 ((size_t)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 ((size_t)tarindex >= targsize)
251 					return (-1);
252 				target[tarindex] |= (pos - Base64);
253 			}
254 			tarindex++;
255 			state = 0;
256 			break;
257 		default:
258 			abort();
259 		}
260 	}
261 
262 	/*
263 	 * We are done decoding Base-64 chars.  Let's see if we ended
264 	 * on a byte boundary, and/or with erroneous trailing characters.
265 	 */
266 
267 	if (ch == Pad64) {		/* We got a pad char. */
268 		ch = *src++;		/* Skip it, get next. */
269 		switch (state) {
270 		case 0:		/* Invalid = in first position */
271 		case 1:		/* Invalid = in second position */
272 			return (-1);
273 
274 		case 2:		/* Valid, means one byte of info */
275 			/* Skip any number of spaces. */
276 			for ((void)NULL; ch != '\0'; ch = *src++)
277 				if (!isspace((unsigned char)ch))
278 					break;
279 			/* Make sure there is another trailing = sign. */
280 			if (ch != Pad64)
281 				return (-1);
282 			ch = *src++;		/* Skip the = */
283 			/* Fall through to "single trailing =" case. */
284 			/* FALLTHROUGH */
285 
286 		case 3:		/* Valid, means two bytes of info */
287 			/*
288 			 * We know this char is an =.  Is there anything but
289 			 * whitespace after it?
290 			 */
291 			for ((void)NULL; ch != '\0'; ch = *src++)
292 				if (!isspace((unsigned char)ch))
293 					return (-1);
294 
295 			/*
296 			 * Now make sure for cases 2 and 3 that the "extra"
297 			 * bits that slopped past the last full byte were
298 			 * zeros.  If we don't check them, they become a
299 			 * subliminal channel.
300 			 */
301 			if (target && target[tarindex] != 0)
302 				return (-1);
303 		}
304 	} else {
305 		/*
306 		 * We ended by seeing the end of the string.  Make sure we
307 		 * have no partial bytes lying around.
308 		 */
309 		if (state != 0)
310 			return (-1);
311 	}
312 
313 	return (tarindex);
314 }
315