xref: /titanic_50/usr/src/cmd/ssh/libopenbsd-compat/common/base64.c (revision 4b22b9337f359bfd063322244f5336cc7c6ffcfa)
1 /*	$OpenBSD: base64.c,v 1.4 2002/01/02 23:00:10 deraadt 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 "includes.h"
46 
47 #if !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)
48 
49 #include <sys/types.h>
50 #include <sys/param.h>
51 #include <sys/socket.h>
52 #include <netinet/in.h>
53 #include <arpa/inet.h>
54 
55 #include <ctype.h>
56 #include <stdio.h>
57 
58 #include <stdlib.h>
59 #include <string.h>
60 
61 #include "base64.h"
62 
63 /* XXX abort illegal in library */
64 #define Assert(Cond) if (!(Cond)) abort()
65 
66 static const char Base64[] =
67 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
68 static const char Pad64 = '=';
69 
70 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
71    The following encoding technique is taken from RFC 1521 by Borenstein
72    and Freed.  It is reproduced here in a slightly edited form for
73    convenience.
74 
75    A 65-character subset of US-ASCII is used, enabling 6 bits to be
76    represented per printable character. (The extra 65th character, "=",
77    is used to signify a special processing function.)
78 
79    The encoding process represents 24-bit groups of input bits as output
80    strings of 4 encoded characters. Proceeding from left to right, a
81    24-bit input group is formed by concatenating 3 8-bit input groups.
82    These 24 bits are then treated as 4 concatenated 6-bit groups, each
83    of which is translated into a single digit in the base64 alphabet.
84 
85    Each 6-bit group is used as an index into an array of 64 printable
86    characters. The character referenced by the index is placed in the
87    output string.
88 
89                          Table 1: The Base64 Alphabet
90 
91       Value Encoding  Value Encoding  Value Encoding  Value Encoding
92           0 A            17 R            34 i            51 z
93           1 B            18 S            35 j            52 0
94           2 C            19 T            36 k            53 1
95           3 D            20 U            37 l            54 2
96           4 E            21 V            38 m            55 3
97           5 F            22 W            39 n            56 4
98           6 G            23 X            40 o            57 5
99           7 H            24 Y            41 p            58 6
100           8 I            25 Z            42 q            59 7
101           9 J            26 a            43 r            60 8
102          10 K            27 b            44 s            61 9
103          11 L            28 c            45 t            62 +
104          12 M            29 d            46 u            63 /
105          13 N            30 e            47 v
106          14 O            31 f            48 w         (pad) =
107          15 P            32 g            49 x
108          16 Q            33 h            50 y
109 
110    Special processing is performed if fewer than 24 bits are available
111    at the end of the data being encoded.  A full encoding quantum is
112    always completed at the end of a quantity.  When fewer than 24 input
113    bits are available in an input group, zero bits are added (on the
114    right) to form an integral number of 6-bit groups.  Padding at the
115    end of the data is performed using the '=' character.
116 
117    Since all base64 input is an integral number of octets, only the
118          -------------------------------------------------
119    following cases can arise:
120 
121        (1) the final quantum of encoding input is an integral
122            multiple of 24 bits; here, the final unit of encoded
123 	   output will be an integral multiple of 4 characters
124 	   with no "=" padding,
125        (2) the final quantum of encoding input is exactly 8 bits;
126            here, the final unit of encoded output will be two
127 	   characters followed by two "=" padding characters, or
128        (3) the final quantum of encoding input is exactly 16 bits;
129            here, the final unit of encoded output will be three
130 	   characters followed by one "=" padding character.
131    */
132 
133 int
134 b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize)
135 {
136 	size_t datalength = 0;
137 	u_char input[3];
138 	u_char output[4];
139 	int i;
140 
141 	while (2 < srclength) {
142 		input[0] = *src++;
143 		input[1] = *src++;
144 		input[2] = *src++;
145 		srclength -= 3;
146 
147 		output[0] = input[0] >> 2;
148 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
149 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
150 		output[3] = input[2] & 0x3f;
151 		Assert(output[0] < 64);
152 		Assert(output[1] < 64);
153 		Assert(output[2] < 64);
154 		Assert(output[3] < 64);
155 
156 		if (datalength + 4 > targsize)
157 			return (-1);
158 		target[datalength++] = Base64[output[0]];
159 		target[datalength++] = Base64[output[1]];
160 		target[datalength++] = Base64[output[2]];
161 		target[datalength++] = Base64[output[3]];
162 	}
163 
164 	/* Now we worry about padding. */
165 	if (0 != srclength) {
166 		/* Get what's left. */
167 		input[0] = input[1] = input[2] = '\0';
168 		for (i = 0; i < srclength; i++)
169 			input[i] = *src++;
170 
171 		output[0] = input[0] >> 2;
172 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
173 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
174 		Assert(output[0] < 64);
175 		Assert(output[1] < 64);
176 		Assert(output[2] < 64);
177 
178 		if (datalength + 4 > targsize)
179 			return (-1);
180 		target[datalength++] = Base64[output[0]];
181 		target[datalength++] = Base64[output[1]];
182 		if (srclength == 1)
183 			target[datalength++] = Pad64;
184 		else
185 			target[datalength++] = Base64[output[2]];
186 		target[datalength++] = Pad64;
187 	}
188 	if (datalength >= targsize)
189 		return (-1);
190 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
191 	return (datalength);
192 }
193 
194 /* skips all whitespace anywhere.
195    converts characters, four at a time, starting at (or after)
196    src from base - 64 numbers into three 8 bit bytes in the target area.
197    it returns the number of data bytes stored at the target, or -1 on error.
198  */
199 
200 int
201 b64_pton(char const *src, u_char *target, size_t targsize)
202 {
203 	int tarindex, state, ch;
204 	char *pos;
205 
206 	state = 0;
207 	tarindex = 0;
208 
209 	while ((ch = *src++) != '\0') {
210 		if (isspace(ch))	/* Skip whitespace anywhere. */
211 			continue;
212 
213 		if (ch == Pad64)
214 			break;
215 
216 		pos = strchr(Base64, ch);
217 		if (pos == 0) 		/* A non-base64 character. */
218 			return (-1);
219 
220 		switch (state) {
221 		case 0:
222 			if (target) {
223 				if (tarindex >= targsize)
224 					return (-1);
225 				target[tarindex] = (pos - Base64) << 2;
226 			}
227 			state = 1;
228 			break;
229 		case 1:
230 			if (target) {
231 				if (tarindex + 1 >= targsize)
232 					return (-1);
233 				target[tarindex]   |=  (pos - Base64) >> 4;
234 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
235 							<< 4 ;
236 			}
237 			tarindex++;
238 			state = 2;
239 			break;
240 		case 2:
241 			if (target) {
242 				if (tarindex + 1 >= targsize)
243 					return (-1);
244 				target[tarindex]   |=  (pos - Base64) >> 2;
245 				target[tarindex+1]  = ((pos - Base64) & 0x03)
246 							<< 6;
247 			}
248 			tarindex++;
249 			state = 3;
250 			break;
251 		case 3:
252 			if (target) {
253 				if (tarindex >= targsize)
254 					return (-1);
255 				target[tarindex] |= (pos - Base64);
256 			}
257 			tarindex++;
258 			state = 0;
259 			break;
260 		}
261 	}
262 
263 	/*
264 	 * We are done decoding Base-64 chars.  Let's see if we ended
265 	 * on a byte boundary, and/or with erroneous trailing characters.
266 	 */
267 
268 	if (ch == Pad64) {		/* We got a pad char. */
269 		ch = *src++;		/* Skip it, get next. */
270 		switch (state) {
271 		case 0:		/* Invalid = in first position */
272 		case 1:		/* Invalid = in second position */
273 			return (-1);
274 
275 		case 2:		/* Valid, means one byte of info */
276 			/* Skip any number of spaces. */
277 			for (; ch != '\0'; ch = *src++)
278 				if (!isspace(ch))
279 					break;
280 			/* Make sure there is another trailing = sign. */
281 			if (ch != Pad64)
282 				return (-1);
283 			ch = *src++;		/* Skip the = */
284 			/* Fall through to "single trailing =" case. */
285 			/* FALLTHROUGH */
286 
287 		case 3:		/* Valid, means two bytes of info */
288 			/*
289 			 * We know this char is an =.  Is there anything but
290 			 * whitespace after it?
291 			 */
292 			for (; ch != '\0'; ch = *src++)
293 				if (!isspace(ch))
294 					return (-1);
295 
296 			/*
297 			 * Now make sure for cases 2 and 3 that the "extra"
298 			 * bits that slopped past the last full byte were
299 			 * zeros.  If we don't check them, they become a
300 			 * subliminal channel.
301 			 */
302 			if (target && target[tarindex] != 0)
303 				return (-1);
304 		}
305 	} else {
306 		/*
307 		 * We ended by seeing the end of the string.  Make sure we
308 		 * have no partial bytes lying around.
309 		 */
310 		if (state != 0)
311 			return (-1);
312 	}
313 
314 	return (tarindex);
315 }
316 
317 #endif /* !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) */
318 
319 #pragma ident	"%Z%%M%	%I%	%E% SMI"
320