1/*	$OpenBSD: base64.c,v 1.5 2006/10/21 09:55:03 otto 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/* OPENBSD ORIGINAL: lib/libc/net/base64.c */
46
47#include "includes.h"
48
49#if (!defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)) || (!defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON))
50
51#include <sys/types.h>
52#include <sys/socket.h>
53#include <netinet/in.h>
54#include <arpa/inet.h>
55
56#include <ctype.h>
57#include <stdio.h>
58
59#include <stdlib.h>
60#include <string.h>
61
62#include "base64.h"
63
64static const char Base64[] =
65	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
66static const char Pad64 = '=';
67
68/* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
69   The following encoding technique is taken from RFC 1521 by Borenstein
70   and Freed.  It is reproduced here in a slightly edited form for
71   convenience.
72
73   A 65-character subset of US-ASCII is used, enabling 6 bits to be
74   represented per printable character. (The extra 65th character, "=",
75   is used to signify a special processing function.)
76
77   The encoding process represents 24-bit groups of input bits as output
78   strings of 4 encoded characters. Proceeding from left to right, a
79   24-bit input group is formed by concatenating 3 8-bit input groups.
80   These 24 bits are then treated as 4 concatenated 6-bit groups, each
81   of which is translated into a single digit in the base64 alphabet.
82
83   Each 6-bit group is used as an index into an array of 64 printable
84   characters. The character referenced by the index is placed in the
85   output string.
86
87                         Table 1: The Base64 Alphabet
88
89      Value Encoding  Value Encoding  Value Encoding  Value Encoding
90          0 A            17 R            34 i            51 z
91          1 B            18 S            35 j            52 0
92          2 C            19 T            36 k            53 1
93          3 D            20 U            37 l            54 2
94          4 E            21 V            38 m            55 3
95          5 F            22 W            39 n            56 4
96          6 G            23 X            40 o            57 5
97          7 H            24 Y            41 p            58 6
98          8 I            25 Z            42 q            59 7
99          9 J            26 a            43 r            60 8
100         10 K            27 b            44 s            61 9
101         11 L            28 c            45 t            62 +
102         12 M            29 d            46 u            63 /
103         13 N            30 e            47 v
104         14 O            31 f            48 w         (pad) =
105         15 P            32 g            49 x
106         16 Q            33 h            50 y
107
108   Special processing is performed if fewer than 24 bits are available
109   at the end of the data being encoded.  A full encoding quantum is
110   always completed at the end of a quantity.  When fewer than 24 input
111   bits are available in an input group, zero bits are added (on the
112   right) to form an integral number of 6-bit groups.  Padding at the
113   end of the data is performed using the '=' character.
114
115   Since all base64 input is an integral number of octets, only the
116         -------------------------------------------------
117   following cases can arise:
118
119       (1) the final quantum of encoding input is an integral
120           multiple of 24 bits; here, the final unit of encoded
121	   output will be an integral multiple of 4 characters
122	   with no "=" padding,
123       (2) the final quantum of encoding input is exactly 8 bits;
124           here, the final unit of encoded output will be two
125	   characters followed by two "=" padding characters, or
126       (3) the final quantum of encoding input is exactly 16 bits;
127           here, the final unit of encoded output will be three
128	   characters followed by one "=" padding character.
129   */
130
131#if !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)
132int
133b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize)
134{
135	size_t datalength = 0;
136	u_char input[3];
137	u_char output[4];
138	u_int i;
139
140	while (2 < srclength) {
141		input[0] = *src++;
142		input[1] = *src++;
143		input[2] = *src++;
144		srclength -= 3;
145
146		output[0] = input[0] >> 2;
147		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
148		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
149		output[3] = input[2] & 0x3f;
150
151		if (datalength + 4 > targsize)
152			return (-1);
153		target[datalength++] = Base64[output[0]];
154		target[datalength++] = Base64[output[1]];
155		target[datalength++] = Base64[output[2]];
156		target[datalength++] = Base64[output[3]];
157	}
158
159	/* Now we worry about padding. */
160	if (0 != srclength) {
161		/* Get what's left. */
162		input[0] = input[1] = input[2] = '\0';
163		for (i = 0; i < srclength; i++)
164			input[i] = *src++;
165
166		output[0] = input[0] >> 2;
167		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
168		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
169
170		if (datalength + 4 > targsize)
171			return (-1);
172		target[datalength++] = Base64[output[0]];
173		target[datalength++] = Base64[output[1]];
174		if (srclength == 1)
175			target[datalength++] = Pad64;
176		else
177			target[datalength++] = Base64[output[2]];
178		target[datalength++] = Pad64;
179	}
180	if (datalength >= targsize)
181		return (-1);
182	target[datalength] = '\0';	/* Returned value doesn't count \0. */
183	return (datalength);
184}
185#endif /* !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) */
186
187#if !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON)
188
189/* skips all whitespace anywhere.
190   converts characters, four at a time, starting at (or after)
191   src from base - 64 numbers into three 8 bit bytes in the target area.
192   it returns the number of data bytes stored at the target, or -1 on error.
193 */
194
195int
196b64_pton(char const *src, u_char *target, size_t targsize)
197{
198	u_int tarindex, state;
199	int ch;
200	char *pos;
201
202	state = 0;
203	tarindex = 0;
204
205	while ((ch = *src++) != '\0') {
206		if (isspace(ch))	/* Skip whitespace anywhere. */
207			continue;
208
209		if (ch == Pad64)
210			break;
211
212		pos = strchr(Base64, ch);
213		if (pos == 0)		/* A non-base64 character. */
214			return (-1);
215
216		switch (state) {
217		case 0:
218			if (target) {
219				if (tarindex >= targsize)
220					return (-1);
221				target[tarindex] = (pos - Base64) << 2;
222			}
223			state = 1;
224			break;
225		case 1:
226			if (target) {
227				if (tarindex + 1 >= targsize)
228					return (-1);
229				target[tarindex]   |=  (pos - Base64) >> 4;
230				target[tarindex+1]  = ((pos - Base64) & 0x0f)
231							<< 4 ;
232			}
233			tarindex++;
234			state = 2;
235			break;
236		case 2:
237			if (target) {
238				if (tarindex + 1 >= targsize)
239					return (-1);
240				target[tarindex]   |=  (pos - Base64) >> 2;
241				target[tarindex+1]  = ((pos - Base64) & 0x03)
242							<< 6;
243			}
244			tarindex++;
245			state = 3;
246			break;
247		case 3:
248			if (target) {
249				if (tarindex >= targsize)
250					return (-1);
251				target[tarindex] |= (pos - Base64);
252			}
253			tarindex++;
254			state = 0;
255			break;
256		}
257	}
258
259	/*
260	 * We are done decoding Base-64 chars.  Let's see if we ended
261	 * on a byte boundary, and/or with erroneous trailing characters.
262	 */
263
264	if (ch == Pad64) {		/* We got a pad char. */
265		ch = *src++;		/* Skip it, get next. */
266		switch (state) {
267		case 0:		/* Invalid = in first position */
268		case 1:		/* Invalid = in second position */
269			return (-1);
270
271		case 2:		/* Valid, means one byte of info */
272			/* Skip any number of spaces. */
273			for (; ch != '\0'; ch = *src++)
274				if (!isspace(ch))
275					break;
276			/* Make sure there is another trailing = sign. */
277			if (ch != Pad64)
278				return (-1);
279			ch = *src++;		/* Skip the = */
280			/* Fall through to "single trailing =" case. */
281			/* FALLTHROUGH */
282
283		case 3:		/* Valid, means two bytes of info */
284			/*
285			 * We know this char is an =.  Is there anything but
286			 * whitespace after it?
287			 */
288			for (; ch != '\0'; ch = *src++)
289				if (!isspace(ch))
290					return (-1);
291
292			/*
293			 * Now make sure for cases 2 and 3 that the "extra"
294			 * bits that slopped past the last full byte were
295			 * zeros.  If we don't check them, they become a
296			 * subliminal channel.
297			 */
298			if (target && 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
313#endif /* !defined(HAVE_B64_PTON) && !defined(HAVE___B64_PTON) */
314#endif
315