1 /*	$OpenBSD: base64.c,v 1.7 2013/12/31 02:32:56 tedu 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 #include <sys/types.h>
46 #include <sys/param.h>
47 #include <sys/socket.h>
48 #include <netinet/in.h>
49 #include <arpa/inet.h>
50 #include <arpa/nameser.h>
51 
52 #include <ctype.h>
53 #include <resolv.h>
54 #include <stdio.h>
55 
56 #include <stdlib.h>
57 #include <string.h>
58 
59 __RCSID("$MirOS: src/lib/libc/net/base64.c,v 1.4 2014/03/05 13:15:11 tg Exp $");
60 
61 extern const uint8_t mbsd_digits_base64[65];
62 static const char Pad64 = '=';
63 
64 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
65    The following encoding technique is taken from RFC 1521 by Borenstein
66    and Freed.  It is reproduced here in a slightly edited form for
67    convenience.
68 
69    A 65-character subset of US-ASCII is used, enabling 6 bits to be
70    represented per printable character. (The extra 65th character, "=",
71    is used to signify a special processing function.)
72 
73    The encoding process represents 24-bit groups of input bits as output
74    strings of 4 encoded characters. Proceeding from left to right, a
75    24-bit input group is formed by concatenating 3 8-bit input groups.
76    These 24 bits are then treated as 4 concatenated 6-bit groups, each
77    of which is translated into a single digit in the base64 alphabet.
78 
79    Each 6-bit group is used as an index into an array of 64 printable
80    characters. The character referenced by the index is placed in the
81    output string.
82 
83                          Table 1: The Base64 Alphabet
84 
85       Value Encoding  Value Encoding  Value Encoding  Value Encoding
86           0 A            17 R            34 i            51 z
87           1 B            18 S            35 j            52 0
88           2 C            19 T            36 k            53 1
89           3 D            20 U            37 l            54 2
90           4 E            21 V            38 m            55 3
91           5 F            22 W            39 n            56 4
92           6 G            23 X            40 o            57 5
93           7 H            24 Y            41 p            58 6
94           8 I            25 Z            42 q            59 7
95           9 J            26 a            43 r            60 8
96          10 K            27 b            44 s            61 9
97          11 L            28 c            45 t            62 +
98          12 M            29 d            46 u            63 /
99          13 N            30 e            47 v
100          14 O            31 f            48 w         (pad) =
101          15 P            32 g            49 x
102          16 Q            33 h            50 y
103 
104    Special processing is performed if fewer than 24 bits are available
105    at the end of the data being encoded.  A full encoding quantum is
106    always completed at the end of a quantity.  When fewer than 24 input
107    bits are available in an input group, zero bits are added (on the
108    right) to form an integral number of 6-bit groups.  Padding at the
109    end of the data is performed using the '=' character.
110 
111    Since all base64 input is an integral number of octets, only the
112          -------------------------------------------------
113    following cases can arise:
114 
115        (1) the final quantum of encoding input is an integral
116            multiple of 24 bits; here, the final unit of encoded
117 	   output will be an integral multiple of 4 characters
118 	   with no "=" padding,
119        (2) the final quantum of encoding input is exactly 8 bits;
120            here, the final unit of encoded output will be two
121 	   characters followed by two "=" padding characters, or
122        (3) the final quantum of encoding input is exactly 16 bits;
123            here, the final unit of encoded output will be three
124 	   characters followed by one "=" padding character.
125    */
126 
127 int
b64_ntop(u_char const * src,size_t srclength,char * target,size_t targsize)128 b64_ntop(u_char const *src, size_t srclength,  char *target, size_t targsize)
129 {
130 	size_t datalength = 0;
131 	u_char input[3];
132 	u_char output[4];
133 	size_t i;
134 
135 	while (2 < srclength) {
136 		input[0] = *src++;
137 		input[1] = *src++;
138 		input[2] = *src++;
139 		srclength -= 3;
140 
141 		output[0] = input[0] >> 2;
142 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
143 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
144 		output[3] = input[2] & 0x3f;
145 
146 		if (datalength + 4 > targsize)
147 			return (-1);
148 		target[datalength++] = mbsd_digits_base64[output[0]];
149 		target[datalength++] = mbsd_digits_base64[output[1]];
150 		target[datalength++] = mbsd_digits_base64[output[2]];
151 		target[datalength++] = mbsd_digits_base64[output[3]];
152 	}
153 
154 	/* Now we worry about padding. */
155 	if (0 != srclength) {
156 		/* Get what's left. */
157 		input[0] = input[1] = input[2] = '\0';
158 		for (i = 0; i < srclength; i++)
159 			input[i] = *src++;
160 
161 		output[0] = input[0] >> 2;
162 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
163 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
164 
165 		if (datalength + 4 > targsize)
166 			return (-1);
167 		target[datalength++] = mbsd_digits_base64[output[0]];
168 		target[datalength++] = mbsd_digits_base64[output[1]];
169 		if (srclength == 1)
170 			target[datalength++] = Pad64;
171 		else
172 			target[datalength++] = mbsd_digits_base64[output[2]];
173 		target[datalength++] = Pad64;
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
b64_pton(char const * src,u_char * target,size_t targsize)188 b64_pton(char const *src, u_char *target, size_t targsize)
189 {
190 	size_t tarindex;
191 	int state, ch;
192 	u_char nextbyte;
193 	const uint8_t *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(mbsd_digits_base64, ch);
206 		if (pos == 0) 		/* 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 - &mbsd_digits_base64[0]) << 2;
215 			}
216 			state = 1;
217 			break;
218 		case 1:
219 			if (target) {
220 				if (tarindex >= targsize)
221 					return (-1);
222 				target[tarindex]   |=  (pos - &mbsd_digits_base64[0]) >> 4;
223 				nextbyte = ((pos - &mbsd_digits_base64[0]) & 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 - &mbsd_digits_base64[0]) >> 2;
237 				nextbyte = ((pos - &mbsd_digits_base64[0]) & 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 - &mbsd_digits_base64[0]);
251 			}
252 			tarindex++;
253 			state = 0;
254 			break;
255 		}
256 	}
257 
258 	/*
259 	 * We are done decoding Base-64 chars.  Let's see if we ended
260 	 * on a byte boundary, and/or with erroneous trailing characters.
261 	 */
262 
263 	if (ch == Pad64) {			/* We got a pad char. */
264 		ch = (unsigned char)*src++;	/* Skip it, get next. */
265 		switch (state) {
266 		case 0:		/* Invalid = in first position */
267 		case 1:		/* Invalid = in second position */
268 			return (-1);
269 
270 		case 2:		/* Valid, means one byte of info */
271 			/* Skip any number of spaces. */
272 			for (; ch != '\0'; ch = (unsigned char)*src++)
273 				if (!isspace(ch))
274 					break;
275 			/* Make sure there is another trailing = sign. */
276 			if (ch != Pad64)
277 				return (-1);
278 			ch = (unsigned char)*src++;		/* Skip the = */
279 			/* Fall through to "single trailing =" case. */
280 			/* FALLTHROUGH */
281 
282 		case 3:		/* Valid, means two bytes of info */
283 			/*
284 			 * We know this char is an =.  Is there anything but
285 			 * whitespace after it?
286 			 */
287 			for (; ch != '\0'; ch = (unsigned char)*src++)
288 				if (!isspace(ch))
289 					return (-1);
290 
291 			/*
292 			 * Now make sure for cases 2 and 3 that the "extra"
293 			 * bits that slopped past the last full byte were
294 			 * zeros.  If we don't check them, they become a
295 			 * subliminal channel.
296 			 */
297 			if (target && tarindex < targsize &&
298 			    target[tarindex] != 0)
299 				return (-1);
300 		}
301 	} else {
302 		/*
303 		 * We ended by seeing the end of the string.  Make sure we
304 		 * have no partial bytes lying around.
305 		 */
306 		if (state != 0)
307 			return (-1);
308 	}
309 
310 	return (tarindex);
311 }
312