1*0Sstevel@tonic-gate /*	$OpenBSD: base64.c,v 1.4 2002/01/02 23:00:10 deraadt Exp $	*/
2*0Sstevel@tonic-gate 
3*0Sstevel@tonic-gate /*
4*0Sstevel@tonic-gate  * Copyright (c) 1996 by Internet Software Consortium.
5*0Sstevel@tonic-gate  *
6*0Sstevel@tonic-gate  * Permission to use, copy, modify, and distribute this software for any
7*0Sstevel@tonic-gate  * purpose with or without fee is hereby granted, provided that the above
8*0Sstevel@tonic-gate  * copyright notice and this permission notice appear in all copies.
9*0Sstevel@tonic-gate  *
10*0Sstevel@tonic-gate  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
11*0Sstevel@tonic-gate  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
12*0Sstevel@tonic-gate  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
13*0Sstevel@tonic-gate  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
14*0Sstevel@tonic-gate  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
15*0Sstevel@tonic-gate  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
16*0Sstevel@tonic-gate  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
17*0Sstevel@tonic-gate  * SOFTWARE.
18*0Sstevel@tonic-gate  */
19*0Sstevel@tonic-gate 
20*0Sstevel@tonic-gate /*
21*0Sstevel@tonic-gate  * Portions Copyright (c) 1995 by International Business Machines, Inc.
22*0Sstevel@tonic-gate  *
23*0Sstevel@tonic-gate  * International Business Machines, Inc. (hereinafter called IBM) grants
24*0Sstevel@tonic-gate  * permission under its copyrights to use, copy, modify, and distribute this
25*0Sstevel@tonic-gate  * Software with or without fee, provided that the above copyright notice and
26*0Sstevel@tonic-gate  * all paragraphs of this notice appear in all copies, and that the name of IBM
27*0Sstevel@tonic-gate  * not be used in connection with the marketing of any product incorporating
28*0Sstevel@tonic-gate  * the Software or modifications thereof, without specific, written prior
29*0Sstevel@tonic-gate  * permission.
30*0Sstevel@tonic-gate  *
31*0Sstevel@tonic-gate  * To the extent it has a right to do so, IBM grants an immunity from suit
32*0Sstevel@tonic-gate  * under its patents, if any, for the use, sale or manufacture of products to
33*0Sstevel@tonic-gate  * the extent that such products are used for performing Domain Name System
34*0Sstevel@tonic-gate  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
35*0Sstevel@tonic-gate  * granted for any product per se or for any other function of any product.
36*0Sstevel@tonic-gate  *
37*0Sstevel@tonic-gate  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
38*0Sstevel@tonic-gate  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
39*0Sstevel@tonic-gate  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
40*0Sstevel@tonic-gate  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
41*0Sstevel@tonic-gate  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
42*0Sstevel@tonic-gate  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
43*0Sstevel@tonic-gate  */
44*0Sstevel@tonic-gate 
45*0Sstevel@tonic-gate #include "includes.h"
46*0Sstevel@tonic-gate 
47*0Sstevel@tonic-gate #if !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP)
48*0Sstevel@tonic-gate 
49*0Sstevel@tonic-gate #include <sys/types.h>
50*0Sstevel@tonic-gate #include <sys/param.h>
51*0Sstevel@tonic-gate #include <sys/socket.h>
52*0Sstevel@tonic-gate #include <netinet/in.h>
53*0Sstevel@tonic-gate #include <arpa/inet.h>
54*0Sstevel@tonic-gate 
55*0Sstevel@tonic-gate #include <ctype.h>
56*0Sstevel@tonic-gate #include <stdio.h>
57*0Sstevel@tonic-gate 
58*0Sstevel@tonic-gate #include <stdlib.h>
59*0Sstevel@tonic-gate #include <string.h>
60*0Sstevel@tonic-gate 
61*0Sstevel@tonic-gate #include "base64.h"
62*0Sstevel@tonic-gate 
63*0Sstevel@tonic-gate /* XXX abort illegal in library */
64*0Sstevel@tonic-gate #define Assert(Cond) if (!(Cond)) abort()
65*0Sstevel@tonic-gate 
66*0Sstevel@tonic-gate static const char Base64[] =
67*0Sstevel@tonic-gate 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
68*0Sstevel@tonic-gate static const char Pad64 = '=';
69*0Sstevel@tonic-gate 
70*0Sstevel@tonic-gate /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
71*0Sstevel@tonic-gate    The following encoding technique is taken from RFC 1521 by Borenstein
72*0Sstevel@tonic-gate    and Freed.  It is reproduced here in a slightly edited form for
73*0Sstevel@tonic-gate    convenience.
74*0Sstevel@tonic-gate 
75*0Sstevel@tonic-gate    A 65-character subset of US-ASCII is used, enabling 6 bits to be
76*0Sstevel@tonic-gate    represented per printable character. (The extra 65th character, "=",
77*0Sstevel@tonic-gate    is used to signify a special processing function.)
78*0Sstevel@tonic-gate 
79*0Sstevel@tonic-gate    The encoding process represents 24-bit groups of input bits as output
80*0Sstevel@tonic-gate    strings of 4 encoded characters. Proceeding from left to right, a
81*0Sstevel@tonic-gate    24-bit input group is formed by concatenating 3 8-bit input groups.
82*0Sstevel@tonic-gate    These 24 bits are then treated as 4 concatenated 6-bit groups, each
83*0Sstevel@tonic-gate    of which is translated into a single digit in the base64 alphabet.
84*0Sstevel@tonic-gate 
85*0Sstevel@tonic-gate    Each 6-bit group is used as an index into an array of 64 printable
86*0Sstevel@tonic-gate    characters. The character referenced by the index is placed in the
87*0Sstevel@tonic-gate    output string.
88*0Sstevel@tonic-gate 
89*0Sstevel@tonic-gate                          Table 1: The Base64 Alphabet
90*0Sstevel@tonic-gate 
91*0Sstevel@tonic-gate       Value Encoding  Value Encoding  Value Encoding  Value Encoding
92*0Sstevel@tonic-gate           0 A            17 R            34 i            51 z
93*0Sstevel@tonic-gate           1 B            18 S            35 j            52 0
94*0Sstevel@tonic-gate           2 C            19 T            36 k            53 1
95*0Sstevel@tonic-gate           3 D            20 U            37 l            54 2
96*0Sstevel@tonic-gate           4 E            21 V            38 m            55 3
97*0Sstevel@tonic-gate           5 F            22 W            39 n            56 4
98*0Sstevel@tonic-gate           6 G            23 X            40 o            57 5
99*0Sstevel@tonic-gate           7 H            24 Y            41 p            58 6
100*0Sstevel@tonic-gate           8 I            25 Z            42 q            59 7
101*0Sstevel@tonic-gate           9 J            26 a            43 r            60 8
102*0Sstevel@tonic-gate          10 K            27 b            44 s            61 9
103*0Sstevel@tonic-gate          11 L            28 c            45 t            62 +
104*0Sstevel@tonic-gate          12 M            29 d            46 u            63 /
105*0Sstevel@tonic-gate          13 N            30 e            47 v
106*0Sstevel@tonic-gate          14 O            31 f            48 w         (pad) =
107*0Sstevel@tonic-gate          15 P            32 g            49 x
108*0Sstevel@tonic-gate          16 Q            33 h            50 y
109*0Sstevel@tonic-gate 
110*0Sstevel@tonic-gate    Special processing is performed if fewer than 24 bits are available
111*0Sstevel@tonic-gate    at the end of the data being encoded.  A full encoding quantum is
112*0Sstevel@tonic-gate    always completed at the end of a quantity.  When fewer than 24 input
113*0Sstevel@tonic-gate    bits are available in an input group, zero bits are added (on the
114*0Sstevel@tonic-gate    right) to form an integral number of 6-bit groups.  Padding at the
115*0Sstevel@tonic-gate    end of the data is performed using the '=' character.
116*0Sstevel@tonic-gate 
117*0Sstevel@tonic-gate    Since all base64 input is an integral number of octets, only the
118*0Sstevel@tonic-gate          -------------------------------------------------
119*0Sstevel@tonic-gate    following cases can arise:
120*0Sstevel@tonic-gate 
121*0Sstevel@tonic-gate        (1) the final quantum of encoding input is an integral
122*0Sstevel@tonic-gate            multiple of 24 bits; here, the final unit of encoded
123*0Sstevel@tonic-gate 	   output will be an integral multiple of 4 characters
124*0Sstevel@tonic-gate 	   with no "=" padding,
125*0Sstevel@tonic-gate        (2) the final quantum of encoding input is exactly 8 bits;
126*0Sstevel@tonic-gate            here, the final unit of encoded output will be two
127*0Sstevel@tonic-gate 	   characters followed by two "=" padding characters, or
128*0Sstevel@tonic-gate        (3) the final quantum of encoding input is exactly 16 bits;
129*0Sstevel@tonic-gate            here, the final unit of encoded output will be three
130*0Sstevel@tonic-gate 	   characters followed by one "=" padding character.
131*0Sstevel@tonic-gate    */
132*0Sstevel@tonic-gate 
133*0Sstevel@tonic-gate int
134*0Sstevel@tonic-gate b64_ntop(u_char const *src, size_t srclength, char *target, size_t targsize)
135*0Sstevel@tonic-gate {
136*0Sstevel@tonic-gate 	size_t datalength = 0;
137*0Sstevel@tonic-gate 	u_char input[3];
138*0Sstevel@tonic-gate 	u_char output[4];
139*0Sstevel@tonic-gate 	int i;
140*0Sstevel@tonic-gate 
141*0Sstevel@tonic-gate 	while (2 < srclength) {
142*0Sstevel@tonic-gate 		input[0] = *src++;
143*0Sstevel@tonic-gate 		input[1] = *src++;
144*0Sstevel@tonic-gate 		input[2] = *src++;
145*0Sstevel@tonic-gate 		srclength -= 3;
146*0Sstevel@tonic-gate 
147*0Sstevel@tonic-gate 		output[0] = input[0] >> 2;
148*0Sstevel@tonic-gate 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
149*0Sstevel@tonic-gate 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
150*0Sstevel@tonic-gate 		output[3] = input[2] & 0x3f;
151*0Sstevel@tonic-gate 		Assert(output[0] < 64);
152*0Sstevel@tonic-gate 		Assert(output[1] < 64);
153*0Sstevel@tonic-gate 		Assert(output[2] < 64);
154*0Sstevel@tonic-gate 		Assert(output[3] < 64);
155*0Sstevel@tonic-gate 
156*0Sstevel@tonic-gate 		if (datalength + 4 > targsize)
157*0Sstevel@tonic-gate 			return (-1);
158*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[0]];
159*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[1]];
160*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[2]];
161*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[3]];
162*0Sstevel@tonic-gate 	}
163*0Sstevel@tonic-gate 
164*0Sstevel@tonic-gate 	/* Now we worry about padding. */
165*0Sstevel@tonic-gate 	if (0 != srclength) {
166*0Sstevel@tonic-gate 		/* Get what's left. */
167*0Sstevel@tonic-gate 		input[0] = input[1] = input[2] = '\0';
168*0Sstevel@tonic-gate 		for (i = 0; i < srclength; i++)
169*0Sstevel@tonic-gate 			input[i] = *src++;
170*0Sstevel@tonic-gate 
171*0Sstevel@tonic-gate 		output[0] = input[0] >> 2;
172*0Sstevel@tonic-gate 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
173*0Sstevel@tonic-gate 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
174*0Sstevel@tonic-gate 		Assert(output[0] < 64);
175*0Sstevel@tonic-gate 		Assert(output[1] < 64);
176*0Sstevel@tonic-gate 		Assert(output[2] < 64);
177*0Sstevel@tonic-gate 
178*0Sstevel@tonic-gate 		if (datalength + 4 > targsize)
179*0Sstevel@tonic-gate 			return (-1);
180*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[0]];
181*0Sstevel@tonic-gate 		target[datalength++] = Base64[output[1]];
182*0Sstevel@tonic-gate 		if (srclength == 1)
183*0Sstevel@tonic-gate 			target[datalength++] = Pad64;
184*0Sstevel@tonic-gate 		else
185*0Sstevel@tonic-gate 			target[datalength++] = Base64[output[2]];
186*0Sstevel@tonic-gate 		target[datalength++] = Pad64;
187*0Sstevel@tonic-gate 	}
188*0Sstevel@tonic-gate 	if (datalength >= targsize)
189*0Sstevel@tonic-gate 		return (-1);
190*0Sstevel@tonic-gate 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
191*0Sstevel@tonic-gate 	return (datalength);
192*0Sstevel@tonic-gate }
193*0Sstevel@tonic-gate 
194*0Sstevel@tonic-gate /* skips all whitespace anywhere.
195*0Sstevel@tonic-gate    converts characters, four at a time, starting at (or after)
196*0Sstevel@tonic-gate    src from base - 64 numbers into three 8 bit bytes in the target area.
197*0Sstevel@tonic-gate    it returns the number of data bytes stored at the target, or -1 on error.
198*0Sstevel@tonic-gate  */
199*0Sstevel@tonic-gate 
200*0Sstevel@tonic-gate int
201*0Sstevel@tonic-gate b64_pton(char const *src, u_char *target, size_t targsize)
202*0Sstevel@tonic-gate {
203*0Sstevel@tonic-gate 	int tarindex, state, ch;
204*0Sstevel@tonic-gate 	char *pos;
205*0Sstevel@tonic-gate 
206*0Sstevel@tonic-gate 	state = 0;
207*0Sstevel@tonic-gate 	tarindex = 0;
208*0Sstevel@tonic-gate 
209*0Sstevel@tonic-gate 	while ((ch = *src++) != '\0') {
210*0Sstevel@tonic-gate 		if (isspace(ch))	/* Skip whitespace anywhere. */
211*0Sstevel@tonic-gate 			continue;
212*0Sstevel@tonic-gate 
213*0Sstevel@tonic-gate 		if (ch == Pad64)
214*0Sstevel@tonic-gate 			break;
215*0Sstevel@tonic-gate 
216*0Sstevel@tonic-gate 		pos = strchr(Base64, ch);
217*0Sstevel@tonic-gate 		if (pos == 0) 		/* A non-base64 character. */
218*0Sstevel@tonic-gate 			return (-1);
219*0Sstevel@tonic-gate 
220*0Sstevel@tonic-gate 		switch (state) {
221*0Sstevel@tonic-gate 		case 0:
222*0Sstevel@tonic-gate 			if (target) {
223*0Sstevel@tonic-gate 				if (tarindex >= targsize)
224*0Sstevel@tonic-gate 					return (-1);
225*0Sstevel@tonic-gate 				target[tarindex] = (pos - Base64) << 2;
226*0Sstevel@tonic-gate 			}
227*0Sstevel@tonic-gate 			state = 1;
228*0Sstevel@tonic-gate 			break;
229*0Sstevel@tonic-gate 		case 1:
230*0Sstevel@tonic-gate 			if (target) {
231*0Sstevel@tonic-gate 				if (tarindex + 1 >= targsize)
232*0Sstevel@tonic-gate 					return (-1);
233*0Sstevel@tonic-gate 				target[tarindex]   |=  (pos - Base64) >> 4;
234*0Sstevel@tonic-gate 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
235*0Sstevel@tonic-gate 							<< 4 ;
236*0Sstevel@tonic-gate 			}
237*0Sstevel@tonic-gate 			tarindex++;
238*0Sstevel@tonic-gate 			state = 2;
239*0Sstevel@tonic-gate 			break;
240*0Sstevel@tonic-gate 		case 2:
241*0Sstevel@tonic-gate 			if (target) {
242*0Sstevel@tonic-gate 				if (tarindex + 1 >= targsize)
243*0Sstevel@tonic-gate 					return (-1);
244*0Sstevel@tonic-gate 				target[tarindex]   |=  (pos - Base64) >> 2;
245*0Sstevel@tonic-gate 				target[tarindex+1]  = ((pos - Base64) & 0x03)
246*0Sstevel@tonic-gate 							<< 6;
247*0Sstevel@tonic-gate 			}
248*0Sstevel@tonic-gate 			tarindex++;
249*0Sstevel@tonic-gate 			state = 3;
250*0Sstevel@tonic-gate 			break;
251*0Sstevel@tonic-gate 		case 3:
252*0Sstevel@tonic-gate 			if (target) {
253*0Sstevel@tonic-gate 				if (tarindex >= targsize)
254*0Sstevel@tonic-gate 					return (-1);
255*0Sstevel@tonic-gate 				target[tarindex] |= (pos - Base64);
256*0Sstevel@tonic-gate 			}
257*0Sstevel@tonic-gate 			tarindex++;
258*0Sstevel@tonic-gate 			state = 0;
259*0Sstevel@tonic-gate 			break;
260*0Sstevel@tonic-gate 		}
261*0Sstevel@tonic-gate 	}
262*0Sstevel@tonic-gate 
263*0Sstevel@tonic-gate 	/*
264*0Sstevel@tonic-gate 	 * We are done decoding Base-64 chars.  Let's see if we ended
265*0Sstevel@tonic-gate 	 * on a byte boundary, and/or with erroneous trailing characters.
266*0Sstevel@tonic-gate 	 */
267*0Sstevel@tonic-gate 
268*0Sstevel@tonic-gate 	if (ch == Pad64) {		/* We got a pad char. */
269*0Sstevel@tonic-gate 		ch = *src++;		/* Skip it, get next. */
270*0Sstevel@tonic-gate 		switch (state) {
271*0Sstevel@tonic-gate 		case 0:		/* Invalid = in first position */
272*0Sstevel@tonic-gate 		case 1:		/* Invalid = in second position */
273*0Sstevel@tonic-gate 			return (-1);
274*0Sstevel@tonic-gate 
275*0Sstevel@tonic-gate 		case 2:		/* Valid, means one byte of info */
276*0Sstevel@tonic-gate 			/* Skip any number of spaces. */
277*0Sstevel@tonic-gate 			for (; ch != '\0'; ch = *src++)
278*0Sstevel@tonic-gate 				if (!isspace(ch))
279*0Sstevel@tonic-gate 					break;
280*0Sstevel@tonic-gate 			/* Make sure there is another trailing = sign. */
281*0Sstevel@tonic-gate 			if (ch != Pad64)
282*0Sstevel@tonic-gate 				return (-1);
283*0Sstevel@tonic-gate 			ch = *src++;		/* Skip the = */
284*0Sstevel@tonic-gate 			/* Fall through to "single trailing =" case. */
285*0Sstevel@tonic-gate 			/* FALLTHROUGH */
286*0Sstevel@tonic-gate 
287*0Sstevel@tonic-gate 		case 3:		/* Valid, means two bytes of info */
288*0Sstevel@tonic-gate 			/*
289*0Sstevel@tonic-gate 			 * We know this char is an =.  Is there anything but
290*0Sstevel@tonic-gate 			 * whitespace after it?
291*0Sstevel@tonic-gate 			 */
292*0Sstevel@tonic-gate 			for (; ch != '\0'; ch = *src++)
293*0Sstevel@tonic-gate 				if (!isspace(ch))
294*0Sstevel@tonic-gate 					return (-1);
295*0Sstevel@tonic-gate 
296*0Sstevel@tonic-gate 			/*
297*0Sstevel@tonic-gate 			 * Now make sure for cases 2 and 3 that the "extra"
298*0Sstevel@tonic-gate 			 * bits that slopped past the last full byte were
299*0Sstevel@tonic-gate 			 * zeros.  If we don't check them, they become a
300*0Sstevel@tonic-gate 			 * subliminal channel.
301*0Sstevel@tonic-gate 			 */
302*0Sstevel@tonic-gate 			if (target && target[tarindex] != 0)
303*0Sstevel@tonic-gate 				return (-1);
304*0Sstevel@tonic-gate 		}
305*0Sstevel@tonic-gate 	} else {
306*0Sstevel@tonic-gate 		/*
307*0Sstevel@tonic-gate 		 * We ended by seeing the end of the string.  Make sure we
308*0Sstevel@tonic-gate 		 * have no partial bytes lying around.
309*0Sstevel@tonic-gate 		 */
310*0Sstevel@tonic-gate 		if (state != 0)
311*0Sstevel@tonic-gate 			return (-1);
312*0Sstevel@tonic-gate 	}
313*0Sstevel@tonic-gate 
314*0Sstevel@tonic-gate 	return (tarindex);
315*0Sstevel@tonic-gate }
316*0Sstevel@tonic-gate 
317*0Sstevel@tonic-gate #endif /* !defined(HAVE_B64_NTOP) && !defined(HAVE___B64_NTOP) */
318*0Sstevel@tonic-gate 
319*0Sstevel@tonic-gate #pragma ident	"%Z%%M%	%I%	%E% SMI"
320