xref: /dragonfly/lib/libc/net/base64.c (revision 8e0b7621)
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/socket.h>
49 
50 #include <netinet/in.h>
51 #include <arpa/inet.h>
52 
53 #include <ctype.h>
54 #include <resolv.h>
55 #include <stdlib.h>
56 #include <string.h>
57 
58 static const char Base64[] =
59 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
60 static const char Pad64 = '=';
61 
62 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
63    The following encoding technique is taken from RFC 1521 by Borenstein
64    and Freed.  It is reproduced here in a slightly edited form for
65    convenience.
66 
67    A 65-character subset of US-ASCII is used, enabling 6 bits to be
68    represented per printable character. (The extra 65th character, "=",
69    is used to signify a special processing function.)
70 
71    The encoding process represents 24-bit groups of input bits as output
72    strings of 4 encoded characters. Proceeding from left to right, a
73    24-bit input group is formed by concatenating 3 8-bit input groups.
74    These 24 bits are then treated as 4 concatenated 6-bit groups, each
75    of which is translated into a single digit in the base64 alphabet.
76 
77    Each 6-bit group is used as an index into an array of 64 printable
78    characters. The character referenced by the index is placed in the
79    output string.
80 
81                          Table 1: The Base64 Alphabet
82 
83       Value Encoding  Value Encoding  Value Encoding  Value Encoding
84           0 A            17 R            34 i            51 z
85           1 B            18 S            35 j            52 0
86           2 C            19 T            36 k            53 1
87           3 D            20 U            37 l            54 2
88           4 E            21 V            38 m            55 3
89           5 F            22 W            39 n            56 4
90           6 G            23 X            40 o            57 5
91           7 H            24 Y            41 p            58 6
92           8 I            25 Z            42 q            59 7
93           9 J            26 a            43 r            60 8
94          10 K            27 b            44 s            61 9
95          11 L            28 c            45 t            62 +
96          12 M            29 d            46 u            63 /
97          13 N            30 e            47 v
98          14 O            31 f            48 w         (pad) =
99          15 P            32 g            49 x
100          16 Q            33 h            50 y
101 
102    Special processing is performed if fewer than 24 bits are available
103    at the end of the data being encoded.  A full encoding quantum is
104    always completed at the end of a quantity.  When fewer than 24 input
105    bits are available in an input group, zero bits are added (on the
106    right) to form an integral number of 6-bit groups.  Padding at the
107    end of the data is performed using the '=' character.
108 
109    Since all base64 input is an integral number of octets, only the
110          -------------------------------------------------
111    following cases can arise:
112 
113        (1) the final quantum of encoding input is an integral
114            multiple of 24 bits; here, the final unit of encoded
115 	   output will be an integral multiple of 4 characters
116 	   with no "=" padding,
117        (2) the final quantum of encoding input is exactly 8 bits;
118            here, the final unit of encoded output will be two
119 	   characters followed by two "=" padding characters, or
120        (3) the final quantum of encoding input is exactly 16 bits;
121            here, the final unit of encoded output will be three
122 	   characters followed by one "=" padding character.
123    */
124 
125 int
126 b64_ntop(const unsigned char *src, size_t srclength, char *target,
127 	 size_t targsize)
128 {
129 	size_t datalength = 0;
130 	unsigned char input[3];
131 	unsigned char output[4];
132 	size_t i;
133 
134 	while (2 < srclength) {
135 		input[0] = *src++;
136 		input[1] = *src++;
137 		input[2] = *src++;
138 		srclength -= 3;
139 
140 		output[0] = input[0] >> 2;
141 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
142 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
143 		output[3] = input[2] & 0x3f;
144 
145 		if (datalength + 4 > targsize)
146 			return (-1);
147 		target[datalength++] = Base64[output[0]];
148 		target[datalength++] = Base64[output[1]];
149 		target[datalength++] = Base64[output[2]];
150 		target[datalength++] = Base64[output[3]];
151 	}
152 
153 	/* Now we worry about padding. */
154 	if (0 != srclength) {
155 		/* Get what's left. */
156 		input[0] = input[1] = input[2] = '\0';
157 		for (i = 0; i < srclength; i++)
158 			input[i] = *src++;
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 
164 		if (datalength + 4 > targsize)
165 			return (-1);
166 		target[datalength++] = Base64[output[0]];
167 		target[datalength++] = Base64[output[1]];
168 		if (srclength == 1)
169 			target[datalength++] = Pad64;
170 		else
171 			target[datalength++] = Base64[output[2]];
172 		target[datalength++] = Pad64;
173 	}
174 
175 	if (datalength >= targsize)
176 		return (-1);
177 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
178 	return (datalength);
179 }
180 
181 /* skips all whitespace anywhere.
182    converts characters, four at a time, starting at (or after)
183    src from base - 64 numbers into three 8 bit bytes in the target area.
184    it returns the number of data bytes stored at the target, or -1 on error.
185  */
186 
187 int
188 b64_pton(const char *src, unsigned char *target, size_t targsize)
189 {
190 	int state;
191 	size_t tarindex;
192 	unsigned char ch, nextbyte;
193 	char *pos;
194 
195 	state = 0;
196 	tarindex = 0;
197 
198 	while ((ch = (unsigned char)*src++) != '\0') {
199 		if (isspace(ch))		/* Skip whitespace anywhere. */
200 			continue;
201 
202 		if (ch == Pad64)
203 			break;
204 
205 		pos = strchr(Base64, ch);
206 		if (pos == NULL)		/* A non-base64 character. */
207 			return (-1);
208 
209 		switch (state) {
210 		case 0:
211 			if (target) {
212 				if (tarindex >= targsize)
213 					return (-1);
214 				target[tarindex] = (pos - Base64) << 2;
215 			}
216 			state = 1;
217 			break;
218 		case 1:
219 			if (target) {
220 				if (tarindex >= targsize)
221 					return (-1);
222 				target[tarindex] |= (pos - Base64) >> 4;
223 				nextbyte = ((pos - Base64) & 0x0f) << 4;
224 				if (tarindex + 1 < targsize)
225 					target[tarindex+1] = nextbyte;
226 				else if (nextbyte)
227 					return (-1);
228 			}
229 			tarindex++;
230 			state = 2;
231 			break;
232 		case 2:
233 			if (target) {
234 				if (tarindex >= targsize)
235 					return (-1);
236 				target[tarindex] |= (pos - Base64) >> 2;
237 				nextbyte = ((pos - Base64) & 0x03) << 6;
238 				if (tarindex + 1 < targsize)
239 					target[tarindex+1] = nextbyte;
240 				else if (nextbyte)
241 					return (-1);
242 			}
243 			tarindex++;
244 			state = 3;
245 			break;
246 		case 3:
247 			if (target) {
248 				if (tarindex >= targsize)
249 					return (-1);
250 				target[tarindex] |= (pos - Base64);
251 			}
252 			tarindex++;
253 			state = 0;
254 			break;
255 		default:	/* Impossible */
256 			abort();
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 = (unsigned char)*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 = (unsigned char)*src++)
275 				if (!isspace(ch))
276 					break;
277 			/* Make sure there is another trailing = sign. */
278 			if (ch != Pad64)
279 				return (-1);
280 			ch = (unsigned char)*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 = (unsigned char)*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 && tarindex < targsize &&
300 			    target[tarindex] != 0)
301 				return (-1);
302 		}
303 	} else {
304 		/*
305 		 * We ended by seeing the end of the string.  Make sure we
306 		 * have no partial bytes lying around.
307 		 */
308 		if (state != 0)
309 			return (-1);
310 	}
311 
312 	return ((int)tarindex);
313 }
314