1*4e179ddaSchristos /* $OpenBSD: base64.c,v 1.8 2015/01/16 16:48:51 deraadt Exp $ */
2*4e179ddaSchristos
3*4e179ddaSchristos /*
4*4e179ddaSchristos * Copyright (c) 1996 by Internet Software Consortium.
5*4e179ddaSchristos *
6*4e179ddaSchristos * Permission to use, copy, modify, and distribute this software for any
7*4e179ddaSchristos * purpose with or without fee is hereby granted, provided that the above
8*4e179ddaSchristos * copyright notice and this permission notice appear in all copies.
9*4e179ddaSchristos *
10*4e179ddaSchristos * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
11*4e179ddaSchristos * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
12*4e179ddaSchristos * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
13*4e179ddaSchristos * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
14*4e179ddaSchristos * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
15*4e179ddaSchristos * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
16*4e179ddaSchristos * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
17*4e179ddaSchristos * SOFTWARE.
18*4e179ddaSchristos */
19*4e179ddaSchristos
20*4e179ddaSchristos /*
21*4e179ddaSchristos * Portions Copyright (c) 1995 by International Business Machines, Inc.
22*4e179ddaSchristos *
23*4e179ddaSchristos * International Business Machines, Inc. (hereinafter called IBM) grants
24*4e179ddaSchristos * permission under its copyrights to use, copy, modify, and distribute this
25*4e179ddaSchristos * Software with or without fee, provided that the above copyright notice and
26*4e179ddaSchristos * all paragraphs of this notice appear in all copies, and that the name of IBM
27*4e179ddaSchristos * not be used in connection with the marketing of any product incorporating
28*4e179ddaSchristos * the Software or modifications thereof, without specific, written prior
29*4e179ddaSchristos * permission.
30*4e179ddaSchristos *
31*4e179ddaSchristos * To the extent it has a right to do so, IBM grants an immunity from suit
32*4e179ddaSchristos * under its patents, if any, for the use, sale or manufacture of products to
33*4e179ddaSchristos * the extent that such products are used for performing Domain Name System
34*4e179ddaSchristos * dynamic updates in TCP/IP networks by means of the Software. No immunity is
35*4e179ddaSchristos * granted for any product per se or for any other function of any product.
36*4e179ddaSchristos *
37*4e179ddaSchristos * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
38*4e179ddaSchristos * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
39*4e179ddaSchristos * PARTICULAR PURPOSE. IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
40*4e179ddaSchristos * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
41*4e179ddaSchristos * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
42*4e179ddaSchristos * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
43*4e179ddaSchristos */
44*4e179ddaSchristos
45*4e179ddaSchristos #include <sys/types.h>
46*4e179ddaSchristos #include <sys/socket.h>
47*4e179ddaSchristos #include <netinet/in.h>
48*4e179ddaSchristos #include <arpa/inet.h>
49*4e179ddaSchristos #include <arpa/nameser.h>
50*4e179ddaSchristos
51*4e179ddaSchristos #include <ctype.h>
52*4e179ddaSchristos #include <resolv.h>
53*4e179ddaSchristos #include <stdio.h>
54*4e179ddaSchristos
55*4e179ddaSchristos #include <stdlib.h>
56*4e179ddaSchristos #include <string.h>
57*4e179ddaSchristos
58*4e179ddaSchristos static const char Base64[] =
59*4e179ddaSchristos "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
60*4e179ddaSchristos static const char Pad64 = '=';
61*4e179ddaSchristos
62*4e179ddaSchristos /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
63*4e179ddaSchristos The following encoding technique is taken from RFC 1521 by Borenstein
64*4e179ddaSchristos and Freed. It is reproduced here in a slightly edited form for
65*4e179ddaSchristos convenience.
66*4e179ddaSchristos
67*4e179ddaSchristos A 65-character subset of US-ASCII is used, enabling 6 bits to be
68*4e179ddaSchristos represented per printable character. (The extra 65th character, "=",
69*4e179ddaSchristos is used to signify a special processing function.)
70*4e179ddaSchristos
71*4e179ddaSchristos The encoding process represents 24-bit groups of input bits as output
72*4e179ddaSchristos strings of 4 encoded characters. Proceeding from left to right, a
73*4e179ddaSchristos 24-bit input group is formed by concatenating 3 8-bit input groups.
74*4e179ddaSchristos These 24 bits are then treated as 4 concatenated 6-bit groups, each
75*4e179ddaSchristos of which is translated into a single digit in the base64 alphabet.
76*4e179ddaSchristos
77*4e179ddaSchristos Each 6-bit group is used as an index into an array of 64 printable
78*4e179ddaSchristos characters. The character referenced by the index is placed in the
79*4e179ddaSchristos output string.
80*4e179ddaSchristos
81*4e179ddaSchristos Table 1: The Base64 Alphabet
82*4e179ddaSchristos
83*4e179ddaSchristos Value Encoding Value Encoding Value Encoding Value Encoding
84*4e179ddaSchristos 0 A 17 R 34 i 51 z
85*4e179ddaSchristos 1 B 18 S 35 j 52 0
86*4e179ddaSchristos 2 C 19 T 36 k 53 1
87*4e179ddaSchristos 3 D 20 U 37 l 54 2
88*4e179ddaSchristos 4 E 21 V 38 m 55 3
89*4e179ddaSchristos 5 F 22 W 39 n 56 4
90*4e179ddaSchristos 6 G 23 X 40 o 57 5
91*4e179ddaSchristos 7 H 24 Y 41 p 58 6
92*4e179ddaSchristos 8 I 25 Z 42 q 59 7
93*4e179ddaSchristos 9 J 26 a 43 r 60 8
94*4e179ddaSchristos 10 K 27 b 44 s 61 9
95*4e179ddaSchristos 11 L 28 c 45 t 62 +
96*4e179ddaSchristos 12 M 29 d 46 u 63 /
97*4e179ddaSchristos 13 N 30 e 47 v
98*4e179ddaSchristos 14 O 31 f 48 w (pad) =
99*4e179ddaSchristos 15 P 32 g 49 x
100*4e179ddaSchristos 16 Q 33 h 50 y
101*4e179ddaSchristos
102*4e179ddaSchristos Special processing is performed if fewer than 24 bits are available
103*4e179ddaSchristos at the end of the data being encoded. A full encoding quantum is
104*4e179ddaSchristos always completed at the end of a quantity. When fewer than 24 input
105*4e179ddaSchristos bits are available in an input group, zero bits are added (on the
106*4e179ddaSchristos right) to form an integral number of 6-bit groups. Padding at the
107*4e179ddaSchristos end of the data is performed using the '=' character.
108*4e179ddaSchristos
109*4e179ddaSchristos Since all base64 input is an integral number of octets, only the
110*4e179ddaSchristos -------------------------------------------------
111*4e179ddaSchristos following cases can arise:
112*4e179ddaSchristos
113*4e179ddaSchristos (1) the final quantum of encoding input is an integral
114*4e179ddaSchristos multiple of 24 bits; here, the final unit of encoded
115*4e179ddaSchristos output will be an integral multiple of 4 characters
116*4e179ddaSchristos with no "=" padding,
117*4e179ddaSchristos (2) the final quantum of encoding input is exactly 8 bits;
118*4e179ddaSchristos here, the final unit of encoded output will be two
119*4e179ddaSchristos characters followed by two "=" padding characters, or
120*4e179ddaSchristos (3) the final quantum of encoding input is exactly 16 bits;
121*4e179ddaSchristos here, the final unit of encoded output will be three
122*4e179ddaSchristos characters followed by one "=" padding character.
123*4e179ddaSchristos */
124*4e179ddaSchristos
125*4e179ddaSchristos int
b64_ntop(src,srclength,target,targsize)126*4e179ddaSchristos b64_ntop(src, srclength, target, targsize)
127*4e179ddaSchristos u_char const *src;
128*4e179ddaSchristos size_t srclength;
129*4e179ddaSchristos char *target;
130*4e179ddaSchristos size_t targsize;
131*4e179ddaSchristos {
132*4e179ddaSchristos size_t datalength = 0;
133*4e179ddaSchristos u_char input[3];
134*4e179ddaSchristos u_char output[4];
135*4e179ddaSchristos int i;
136*4e179ddaSchristos
137*4e179ddaSchristos while (2 < srclength) {
138*4e179ddaSchristos input[0] = *src++;
139*4e179ddaSchristos input[1] = *src++;
140*4e179ddaSchristos input[2] = *src++;
141*4e179ddaSchristos srclength -= 3;
142*4e179ddaSchristos
143*4e179ddaSchristos output[0] = input[0] >> 2;
144*4e179ddaSchristos output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
145*4e179ddaSchristos output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
146*4e179ddaSchristos output[3] = input[2] & 0x3f;
147*4e179ddaSchristos
148*4e179ddaSchristos if (datalength + 4 > targsize)
149*4e179ddaSchristos return (-1);
150*4e179ddaSchristos target[datalength++] = Base64[output[0]];
151*4e179ddaSchristos target[datalength++] = Base64[output[1]];
152*4e179ddaSchristos target[datalength++] = Base64[output[2]];
153*4e179ddaSchristos target[datalength++] = Base64[output[3]];
154*4e179ddaSchristos }
155*4e179ddaSchristos
156*4e179ddaSchristos /* Now we worry about padding. */
157*4e179ddaSchristos if (0 != srclength) {
158*4e179ddaSchristos /* Get what's left. */
159*4e179ddaSchristos input[0] = input[1] = input[2] = '\0';
160*4e179ddaSchristos for (i = 0; i < srclength; i++)
161*4e179ddaSchristos input[i] = *src++;
162*4e179ddaSchristos
163*4e179ddaSchristos output[0] = input[0] >> 2;
164*4e179ddaSchristos output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
165*4e179ddaSchristos output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
166*4e179ddaSchristos
167*4e179ddaSchristos if (datalength + 4 > targsize)
168*4e179ddaSchristos return (-1);
169*4e179ddaSchristos target[datalength++] = Base64[output[0]];
170*4e179ddaSchristos target[datalength++] = Base64[output[1]];
171*4e179ddaSchristos if (srclength == 1)
172*4e179ddaSchristos target[datalength++] = Pad64;
173*4e179ddaSchristos else
174*4e179ddaSchristos target[datalength++] = Base64[output[2]];
175*4e179ddaSchristos target[datalength++] = Pad64;
176*4e179ddaSchristos }
177*4e179ddaSchristos if (datalength >= targsize)
178*4e179ddaSchristos return (-1);
179*4e179ddaSchristos target[datalength] = '\0'; /* Returned value doesn't count \0. */
180*4e179ddaSchristos return (datalength);
181*4e179ddaSchristos }
182*4e179ddaSchristos
183*4e179ddaSchristos /* skips all whitespace anywhere.
184*4e179ddaSchristos converts characters, four at a time, starting at (or after)
185*4e179ddaSchristos src from base - 64 numbers into three 8 bit bytes in the target area.
186*4e179ddaSchristos it returns the number of data bytes stored at the target, or -1 on error.
187*4e179ddaSchristos */
188*4e179ddaSchristos
189*4e179ddaSchristos int
b64_pton(src,target,targsize)190*4e179ddaSchristos b64_pton(src, target, targsize)
191*4e179ddaSchristos char const *src;
192*4e179ddaSchristos u_char *target;
193*4e179ddaSchristos size_t targsize;
194*4e179ddaSchristos {
195*4e179ddaSchristos int tarindex, state, ch;
196*4e179ddaSchristos u_char nextbyte;
197*4e179ddaSchristos char *pos;
198*4e179ddaSchristos
199*4e179ddaSchristos state = 0;
200*4e179ddaSchristos tarindex = 0;
201*4e179ddaSchristos
202*4e179ddaSchristos while ((ch = (unsigned char)*src++) != '\0') {
203*4e179ddaSchristos if (isspace(ch)) /* Skip whitespace anywhere. */
204*4e179ddaSchristos continue;
205*4e179ddaSchristos
206*4e179ddaSchristos if (ch == Pad64)
207*4e179ddaSchristos break;
208*4e179ddaSchristos
209*4e179ddaSchristos pos = strchr(Base64, ch);
210*4e179ddaSchristos if (pos == 0) /* A non-base64 character. */
211*4e179ddaSchristos return (-1);
212*4e179ddaSchristos
213*4e179ddaSchristos switch (state) {
214*4e179ddaSchristos case 0:
215*4e179ddaSchristos if (target) {
216*4e179ddaSchristos if (tarindex >= targsize)
217*4e179ddaSchristos return (-1);
218*4e179ddaSchristos target[tarindex] = (pos - Base64) << 2;
219*4e179ddaSchristos }
220*4e179ddaSchristos state = 1;
221*4e179ddaSchristos break;
222*4e179ddaSchristos case 1:
223*4e179ddaSchristos if (target) {
224*4e179ddaSchristos if (tarindex >= targsize)
225*4e179ddaSchristos return (-1);
226*4e179ddaSchristos target[tarindex] |= (pos - Base64) >> 4;
227*4e179ddaSchristos nextbyte = ((pos - Base64) & 0x0f) << 4;
228*4e179ddaSchristos if (tarindex + 1 < targsize)
229*4e179ddaSchristos target[tarindex+1] = nextbyte;
230*4e179ddaSchristos else if (nextbyte)
231*4e179ddaSchristos return (-1);
232*4e179ddaSchristos }
233*4e179ddaSchristos tarindex++;
234*4e179ddaSchristos state = 2;
235*4e179ddaSchristos break;
236*4e179ddaSchristos case 2:
237*4e179ddaSchristos if (target) {
238*4e179ddaSchristos if (tarindex >= targsize)
239*4e179ddaSchristos return (-1);
240*4e179ddaSchristos target[tarindex] |= (pos - Base64) >> 2;
241*4e179ddaSchristos nextbyte = ((pos - Base64) & 0x03) << 6;
242*4e179ddaSchristos if (tarindex + 1 < targsize)
243*4e179ddaSchristos target[tarindex+1] = nextbyte;
244*4e179ddaSchristos else if (nextbyte)
245*4e179ddaSchristos return (-1);
246*4e179ddaSchristos }
247*4e179ddaSchristos tarindex++;
248*4e179ddaSchristos state = 3;
249*4e179ddaSchristos break;
250*4e179ddaSchristos case 3:
251*4e179ddaSchristos if (target) {
252*4e179ddaSchristos if (tarindex >= targsize)
253*4e179ddaSchristos return (-1);
254*4e179ddaSchristos target[tarindex] |= (pos - Base64);
255*4e179ddaSchristos }
256*4e179ddaSchristos tarindex++;
257*4e179ddaSchristos state = 0;
258*4e179ddaSchristos break;
259*4e179ddaSchristos }
260*4e179ddaSchristos }
261*4e179ddaSchristos
262*4e179ddaSchristos /*
263*4e179ddaSchristos * We are done decoding Base-64 chars. Let's see if we ended
264*4e179ddaSchristos * on a byte boundary, and/or with erroneous trailing characters.
265*4e179ddaSchristos */
266*4e179ddaSchristos
267*4e179ddaSchristos if (ch == Pad64) { /* We got a pad char. */
268*4e179ddaSchristos ch = (unsigned char)*src++; /* Skip it, get next. */
269*4e179ddaSchristos switch (state) {
270*4e179ddaSchristos case 0: /* Invalid = in first position */
271*4e179ddaSchristos case 1: /* Invalid = in second position */
272*4e179ddaSchristos return (-1);
273*4e179ddaSchristos
274*4e179ddaSchristos case 2: /* Valid, means one byte of info */
275*4e179ddaSchristos /* Skip any number of spaces. */
276*4e179ddaSchristos for (; ch != '\0'; ch = (unsigned char)*src++)
277*4e179ddaSchristos if (!isspace(ch))
278*4e179ddaSchristos break;
279*4e179ddaSchristos /* Make sure there is another trailing = sign. */
280*4e179ddaSchristos if (ch != Pad64)
281*4e179ddaSchristos return (-1);
282*4e179ddaSchristos ch = (unsigned char)*src++; /* Skip the = */
283*4e179ddaSchristos /* Fall through to "single trailing =" case. */
284*4e179ddaSchristos /* FALLTHROUGH */
285*4e179ddaSchristos
286*4e179ddaSchristos case 3: /* Valid, means two bytes of info */
287*4e179ddaSchristos /*
288*4e179ddaSchristos * We know this char is an =. Is there anything but
289*4e179ddaSchristos * whitespace after it?
290*4e179ddaSchristos */
291*4e179ddaSchristos for (; ch != '\0'; ch = (unsigned char)*src++)
292*4e179ddaSchristos if (!isspace(ch))
293*4e179ddaSchristos return (-1);
294*4e179ddaSchristos
295*4e179ddaSchristos /*
296*4e179ddaSchristos * Now make sure for cases 2 and 3 that the "extra"
297*4e179ddaSchristos * bits that slopped past the last full byte were
298*4e179ddaSchristos * zeros. If we don't check them, they become a
299*4e179ddaSchristos * subliminal channel.
300*4e179ddaSchristos */
301*4e179ddaSchristos if (target && tarindex < targsize &&
302*4e179ddaSchristos target[tarindex] != 0)
303*4e179ddaSchristos return (-1);
304*4e179ddaSchristos }
305*4e179ddaSchristos } else {
306*4e179ddaSchristos /*
307*4e179ddaSchristos * We ended by seeing the end of the string. Make sure we
308*4e179ddaSchristos * have no partial bytes lying around.
309*4e179ddaSchristos */
310*4e179ddaSchristos if (state != 0)
311*4e179ddaSchristos return (-1);
312*4e179ddaSchristos }
313*4e179ddaSchristos
314*4e179ddaSchristos return (tarindex);
315*4e179ddaSchristos }
316