xref: /freebsd/contrib/ldns/compat/b64_pton.c (revision 7d99ab9fd0cc2c1ce2ecef0ed6d0672c2a50b0cb)
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 #include <ldns/config.h>
43 #ifndef HAVE_B64_PTON
44 
45 #include <sys/types.h>
46 #include <sys/param.h>
47 #ifdef HAVE_SYS_SOCKET_H
48 #include <sys/socket.h>
49 #endif
50 
51 #ifdef HAVE_NETINET_IN_H
52 #include <netinet/in.h>
53 #endif
54 #ifdef HAVE_ARPA_INET_H
55 #include <arpa/inet.h>
56 #endif
57 
58 #include <ctype.h>
59 #include <stdio.h>
60 #include <stdlib.h>
61 #include <string.h>
62 
63 #include <ldns/util.h>
64 
65 static const char Base64[] =
66 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
67 static const char Pad64 = '=';
68 
69 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
70    The following encoding technique is taken from RFC 1521 by Borenstein
71    and Freed.  It is reproduced here in a slightly edited form for
72    convenience.
73 
74    A 65-character subset of US-ASCII is used, enabling 6 bits to be
75    represented per printable character. (The extra 65th character, "=",
76    is used to signify a special processing function.)
77 
78    The encoding process represents 24-bit groups of input bits as output
79    strings of 4 encoded characters. Proceeding from left to right, a
80    24-bit input group is formed by concatenating 3 8-bit input groups.
81    These 24 bits are then treated as 4 concatenated 6-bit groups, each
82    of which is translated into a single digit in the base64 alphabet.
83 
84    Each 6-bit group is used as an index into an array of 64 printable
85    characters. The character referenced by the index is placed in the
86    output string.
87 
88                          Table 1: The Base64 Alphabet
89 
90       Value Encoding  Value Encoding  Value Encoding  Value Encoding
91           0 A            17 R            34 i            51 z
92           1 B            18 S            35 j            52 0
93           2 C            19 T            36 k            53 1
94           3 D            20 U            37 l            54 2
95           4 E            21 V            38 m            55 3
96           5 F            22 W            39 n            56 4
97           6 G            23 X            40 o            57 5
98           7 H            24 Y            41 p            58 6
99           8 I            25 Z            42 q            59 7
100           9 J            26 a            43 r            60 8
101          10 K            27 b            44 s            61 9
102          11 L            28 c            45 t            62 +
103          12 M            29 d            46 u            63 /
104          13 N            30 e            47 v
105          14 O            31 f            48 w         (pad) =
106          15 P            32 g            49 x
107          16 Q            33 h            50 y
108 
109    Special processing is performed if fewer than 24 bits are available
110    at the end of the data being encoded.  A full encoding quantum is
111    always completed at the end of a quantity.  When fewer than 24 input
112    bits are available in an input group, zero bits are added (on the
113    right) to form an integral number of 6-bit groups.  Padding at the
114    end of the data is performed using the '=' character.
115 
116    Since all base64 input is an integral number of octets, only the
117          -------------------------------------------------
118    following cases can arise:
119 
120        (1) the final quantum of encoding input is an integral
121            multiple of 24 bits; here, the final unit of encoded
122 	   output will be an integral multiple of 4 characters
123 	   with no "=" padding,
124        (2) the final quantum of encoding input is exactly 8 bits;
125            here, the final unit of encoded output will be two
126 	   characters followed by two "=" padding characters, or
127        (3) the final quantum of encoding input is exactly 16 bits;
128            here, the final unit of encoded output will be three
129 	   characters followed by one "=" padding character.
130    */
131 
132 /* skips all whitespace anywhere.
133    converts characters, four at a time, starting at (or after)
134    src from base - 64 numbers into three 8 bit bytes in the target area.
135    it returns the number of data bytes stored at the target, or -1 on error.
136  */
137 
138 int
139 ldns_b64_pton(char const *src, uint8_t *target, size_t targsize)
140 {
141 	int tarindex, state, ch;
142 	char *pos;
143 
144 	state = 0;
145 	tarindex = 0;
146 
147 	if (strlen(src) == 0) {
148 		return 0;
149 	}
150 
151 	while ((ch = *src++) != '\0') {
152 		if (isspace((unsigned char)ch))        /* Skip whitespace anywhere. */
153 			continue;
154 
155 		if (ch == Pad64)
156 			break;
157 
158 		pos = strchr(Base64, ch);
159 		if (pos == 0) {
160 			/* A non-base64 character. */
161 			return (-1);
162 		}
163 
164 		switch (state) {
165 		case 0:
166 			if (target) {
167 				if ((size_t)tarindex >= targsize)
168 					return (-1);
169 				target[tarindex] = (pos - Base64) << 2;
170 			}
171 			state = 1;
172 			break;
173 		case 1:
174 			if (target) {
175 				if ((size_t)tarindex + 1 >= targsize)
176 					return (-1);
177 				target[tarindex]   |=  (pos - Base64) >> 4;
178 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
179 							<< 4 ;
180 			}
181 			tarindex++;
182 			state = 2;
183 			break;
184 		case 2:
185 			if (target) {
186 				if ((size_t)tarindex + 1 >= targsize)
187 					return (-1);
188 				target[tarindex]   |=  (pos - Base64) >> 2;
189 				target[tarindex+1]  = ((pos - Base64) & 0x03)
190 							<< 6;
191 			}
192 			tarindex++;
193 			state = 3;
194 			break;
195 		case 3:
196 			if (target) {
197 				if ((size_t)tarindex >= targsize)
198 					return (-1);
199 				target[tarindex] |= (pos - Base64);
200 			}
201 			tarindex++;
202 			state = 0;
203 			break;
204 		default:
205 			abort();
206 		}
207 	}
208 
209 	/*
210 	 * We are done decoding Base-64 chars.  Let's see if we ended
211 	 * on a byte boundary, and/or with erroneous trailing characters.
212 	 */
213 
214 	if (ch == Pad64) {		/* We got a pad char. */
215 		ch = *src++;		/* Skip it, get next. */
216 		switch (state) {
217 		case 0:		/* Invalid = in first position */
218 		case 1:		/* Invalid = in second position */
219 			return (-1);
220 
221 		case 2:		/* Valid, means one byte of info */
222 			/* Skip any number of spaces. */
223 			for ((void)NULL; ch != '\0'; ch = *src++)
224 				if (!isspace((unsigned char)ch))
225 					break;
226 			/* Make sure there is another trailing = sign. */
227 			if (ch != Pad64)
228 				return (-1);
229 			ch = *src++;		/* Skip the = */
230 			/* Fall through to "single trailing =" case. */
231 			/* FALLTHROUGH */
232 
233 		case 3:		/* Valid, means two bytes of info */
234 			/*
235 			 * We know this char is an =.  Is there anything but
236 			 * whitespace after it?
237 			 */
238 			for ((void)NULL; ch != '\0'; ch = *src++)
239 				if (!isspace((unsigned char)ch))
240 					return (-1);
241 
242 			/*
243 			 * Now make sure for cases 2 and 3 that the "extra"
244 			 * bits that slopped past the last full byte were
245 			 * zeros.  If we don't check them, they become a
246 			 * subliminal channel.
247 			 */
248 			if (target && target[tarindex] != 0)
249 				return (-1);
250 		}
251 	} else {
252 		/*
253 		 * We ended by seeing the end of the string.  Make sure we
254 		 * have no partial bytes lying around.
255 		 */
256 		if (state != 0)
257 			return (-1);
258 	}
259 
260 	return (tarindex);
261 }
262 
263 #endif /* !HAVE_B64_PTON */
264