1 /*	$OpenBSD: base64.c,v 1.15 2021/10/25 14:41:09 jca 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/socket.h>
47 #include <netinet/in.h>
48 #include <arpa/inet.h>
49 
50 #include <ctype.h>
51 #include <resolv.h>
52 
53 #include <stdlib.h>
54 #include <string.h>
55 
56 static const char Base64[] =
57 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
58 static const char Pad64 = '=';
59 
60 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
61    The following encoding technique is taken from RFC 1521 by Borenstein
62    and Freed.  It is reproduced here in a slightly edited form for
63    convenience.
64 
65    A 65-character subset of US-ASCII is used, enabling 6 bits to be
66    represented per printable character. (The extra 65th character, "=",
67    is used to signify a special processing function.)
68 
69    The encoding process represents 24-bit groups of input bits as output
70    strings of 4 encoded characters. Proceeding from left to right, a
71    24-bit input group is formed by concatenating 3 8-bit input groups.
72    These 24 bits are then treated as 4 concatenated 6-bit groups, each
73    of which is translated into a single digit in the base64 alphabet.
74 
75    Each 6-bit group is used as an index into an array of 64 printable
76    characters. The character referenced by the index is placed in the
77    output string.
78 
79                          Table 1: The Base64 Alphabet
80 
81       Value Encoding  Value Encoding  Value Encoding  Value Encoding
82           0 A            17 R            34 i            51 z
83           1 B            18 S            35 j            52 0
84           2 C            19 T            36 k            53 1
85           3 D            20 U            37 l            54 2
86           4 E            21 V            38 m            55 3
87           5 F            22 W            39 n            56 4
88           6 G            23 X            40 o            57 5
89           7 H            24 Y            41 p            58 6
90           8 I            25 Z            42 q            59 7
91           9 J            26 a            43 r            60 8
92          10 K            27 b            44 s            61 9
93          11 L            28 c            45 t            62 +
94          12 M            29 d            46 u            63 /
95          13 N            30 e            47 v
96          14 O            31 f            48 w         (pad) =
97          15 P            32 g            49 x
98          16 Q            33 h            50 y
99 
100    Special processing is performed if fewer than 24 bits are available
101    at the end of the data being encoded.  A full encoding quantum is
102    always completed at the end of a quantity.  When fewer than 24 input
103    bits are available in an input group, zero bits are added (on the
104    right) to form an integral number of 6-bit groups.  Padding at the
105    end of the data is performed using the '=' character.
106 
107    Since all base64 input is an integral number of octets, only the
108          -------------------------------------------------
109    following cases can arise:
110 
111        (1) the final quantum of encoding input is an integral
112            multiple of 24 bits; here, the final unit of encoded
113 	   output will be an integral multiple of 4 characters
114 	   with no "=" padding,
115        (2) the final quantum of encoding input is exactly 8 bits;
116            here, the final unit of encoded output will be two
117 	   characters followed by two "=" padding characters, or
118        (3) the final quantum of encoding input is exactly 16 bits;
119            here, the final unit of encoded output will be three
120 	   characters followed by one "=" padding character.
121    */
122 
123 int
b64_ntop(unsigned char const * src,size_t srclength,char * target,size_t targsize)124 b64_ntop(unsigned char const *src, size_t srclength, char *target,
125     size_t targsize)
126 {
127 	size_t datalength = 0;
128 	unsigned char input[3];
129 	unsigned char output[4];
130 	int i;
131 
132 	while (2 < srclength) {
133 		input[0] = *src++;
134 		input[1] = *src++;
135 		input[2] = *src++;
136 		srclength -= 3;
137 
138 		output[0] = input[0] >> 2;
139 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
140 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
141 		output[3] = input[2] & 0x3f;
142 
143 		if (datalength + 4 > targsize)
144 			return (-1);
145 		target[datalength++] = Base64[output[0]];
146 		target[datalength++] = Base64[output[1]];
147 		target[datalength++] = Base64[output[2]];
148 		target[datalength++] = Base64[output[3]];
149 	}
150 
151 	/* Now we worry about padding. */
152 	if (0 != srclength) {
153 		/* Get what's left. */
154 		input[0] = input[1] = input[2] = '\0';
155 		for (i = 0; i < srclength; i++)
156 			input[i] = *src++;
157 
158 		output[0] = input[0] >> 2;
159 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
160 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
161 
162 		if (datalength + 4 > targsize)
163 			return (-1);
164 		target[datalength++] = Base64[output[0]];
165 		target[datalength++] = Base64[output[1]];
166 		if (srclength == 1)
167 			target[datalength++] = Pad64;
168 		else
169 			target[datalength++] = Base64[output[2]];
170 		target[datalength++] = Pad64;
171 	}
172 	if (datalength >= targsize)
173 		return (-1);
174 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
175 	return (datalength);
176 }
177 
178 /* skips all whitespace anywhere.
179    converts characters, four at a time, starting at (or after)
180    src from base - 64 numbers into three 8 bit bytes in the target area.
181    it returns the number of data bytes stored at the target, or -1 on error.
182  */
183 
184 int
b64_pton(char const * src,unsigned char * target,size_t targsize)185 b64_pton(char const *src, unsigned char *target, size_t targsize)
186 {
187 	int tarindex, state, ch;
188 	unsigned char nextbyte;
189 	char *pos;
190 
191 	state = 0;
192 	tarindex = 0;
193 
194 	while ((ch = (unsigned char)*src++) != '\0') {
195 		if (isspace(ch))	/* Skip whitespace anywhere. */
196 			continue;
197 
198 		if (ch == Pad64)
199 			break;
200 
201 		pos = strchr(Base64, ch);
202 		if (pos == 0)		/* A non-base64 character. */
203 			return (-1);
204 
205 		switch (state) {
206 		case 0:
207 			if (target) {
208 				if (tarindex >= targsize)
209 					return (-1);
210 				target[tarindex] = (pos - Base64) << 2;
211 			}
212 			state = 1;
213 			break;
214 		case 1:
215 			if (target) {
216 				if (tarindex >= targsize)
217 					return (-1);
218 				target[tarindex]   |=  (pos - Base64) >> 4;
219 				nextbyte = ((pos - Base64) & 0x0f) << 4;
220 				if (tarindex + 1 < targsize)
221 					target[tarindex+1] = nextbyte;
222 				else if (nextbyte)
223 					return (-1);
224 			}
225 			tarindex++;
226 			state = 2;
227 			break;
228 		case 2:
229 			if (target) {
230 				if (tarindex >= targsize)
231 					return (-1);
232 				target[tarindex]   |=  (pos - Base64) >> 2;
233 				nextbyte = ((pos - Base64) & 0x03) << 6;
234 				if (tarindex + 1 < targsize)
235 					target[tarindex+1] = nextbyte;
236 				else if (nextbyte)
237 					return (-1);
238 			}
239 			tarindex++;
240 			state = 3;
241 			break;
242 		case 3:
243 			if (target) {
244 				if (tarindex >= targsize)
245 					return (-1);
246 				target[tarindex] |= (pos - Base64);
247 			}
248 			tarindex++;
249 			state = 0;
250 			break;
251 		}
252 	}
253 
254 	/*
255 	 * We are done decoding Base-64 chars.  Let's see if we ended
256 	 * on a byte boundary, and/or with erroneous trailing characters.
257 	 */
258 
259 	if (ch == Pad64) {			/* We got a pad char. */
260 		ch = (unsigned char)*src++;	/* Skip it, get next. */
261 		switch (state) {
262 		case 0:		/* Invalid = in first position */
263 		case 1:		/* Invalid = in second position */
264 			return (-1);
265 
266 		case 2:		/* Valid, means one byte of info */
267 			/* Skip any number of spaces. */
268 			for (; ch != '\0'; ch = (unsigned char)*src++)
269 				if (!isspace(ch))
270 					break;
271 			/* Make sure there is another trailing = sign. */
272 			if (ch != Pad64)
273 				return (-1);
274 			ch = (unsigned char)*src++;		/* Skip the = */
275 			/* Fall through to "single trailing =" case. */
276 			/* FALLTHROUGH */
277 
278 		case 3:		/* Valid, means two bytes of info */
279 			/*
280 			 * We know this char is an =.  Is there anything but
281 			 * whitespace after it?
282 			 */
283 			for (; ch != '\0'; ch = (unsigned char)*src++)
284 				if (!isspace(ch))
285 					return (-1);
286 
287 			/*
288 			 * Now make sure for cases 2 and 3 that the "extra"
289 			 * bits that slopped past the last full byte were
290 			 * zeros.  If we don't check them, they become a
291 			 * subliminal channel.
292 			 */
293 			if (target && tarindex < targsize &&
294 			    target[tarindex] != 0)
295 				return (-1);
296 		}
297 	} else {
298 		/*
299 		 * We ended by seeing the end of the string.  Make sure we
300 		 * have no partial bytes lying around.
301 		 */
302 		if (state != 0)
303 			return (-1);
304 	}
305 
306 	return (tarindex);
307 }
308