xref: /netbsd-src/external/bsd/nsd/dist/compat/b64_pton.c (revision f3d63a561b22e5536a0779632985057ed3d766aa)
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 <config.h>
43 
44 #include <sys/types.h>
45 #include <sys/param.h>
46 #include <sys/socket.h>
47 
48 #include <netinet/in.h>
49 #include <arpa/inet.h>
50 
51 #include <ctype.h>
52 #include <stdio.h>
53 #include <stdlib.h>
54 #include <string.h>
55 
56 #define Assert(Cond) if (!(Cond)) abort()
57 
58 static const char Base64[] =
59 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
60 static const char Pad64 = '=';
61 
62 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
63    The following encoding technique is taken from RFC 1521 by Borenstein
64    and Freed.  It is reproduced here in a slightly edited form for
65    convenience.
66 
67    A 65-character subset of US-ASCII is used, enabling 6 bits to be
68    represented per printable character. (The extra 65th character, "=",
69    is used to signify a special processing function.)
70 
71    The encoding process represents 24-bit groups of input bits as output
72    strings of 4 encoded characters. Proceeding from left to right, a
73    24-bit input group is formed by concatenating 3 8-bit input groups.
74    These 24 bits are then treated as 4 concatenated 6-bit groups, each
75    of which is translated into a single digit in the base64 alphabet.
76 
77    Each 6-bit group is used as an index into an array of 64 printable
78    characters. The character referenced by the index is placed in the
79    output string.
80 
81                          Table 1: The Base64 Alphabet
82 
83       Value Encoding  Value Encoding  Value Encoding  Value Encoding
84           0 A            17 R            34 i            51 z
85           1 B            18 S            35 j            52 0
86           2 C            19 T            36 k            53 1
87           3 D            20 U            37 l            54 2
88           4 E            21 V            38 m            55 3
89           5 F            22 W            39 n            56 4
90           6 G            23 X            40 o            57 5
91           7 H            24 Y            41 p            58 6
92           8 I            25 Z            42 q            59 7
93           9 J            26 a            43 r            60 8
94          10 K            27 b            44 s            61 9
95          11 L            28 c            45 t            62 +
96          12 M            29 d            46 u            63 /
97          13 N            30 e            47 v
98          14 O            31 f            48 w         (pad) =
99          15 P            32 g            49 x
100          16 Q            33 h            50 y
101 
102    Special processing is performed if fewer than 24 bits are available
103    at the end of the data being encoded.  A full encoding quantum is
104    always completed at the end of a quantity.  When fewer than 24 input
105    bits are available in an input group, zero bits are added (on the
106    right) to form an integral number of 6-bit groups.  Padding at the
107    end of the data is performed using the '=' character.
108 
109    Since all base64 input is an integral number of octets, only the
110          -------------------------------------------------
111    following cases can arise:
112 
113        (1) the final quantum of encoding input is an integral
114            multiple of 24 bits; here, the final unit of encoded
115 	   output will be an integral multiple of 4 characters
116 	   with no "=" padding,
117        (2) the final quantum of encoding input is exactly 8 bits;
118            here, the final unit of encoded output will be two
119 	   characters followed by two "=" padding characters, or
120        (3) the final quantum of encoding input is exactly 16 bits;
121            here, the final unit of encoded output will be three
122 	   characters followed by one "=" padding character.
123    */
124 
125 /* skips all whitespace anywhere.
126    converts characters, four at a time, starting at (or after)
127    src from base - 64 numbers into three 8 bit bytes in the target area.
128    it returns the number of data bytes stored at the target, or -1 on error.
129  */
130 
131 static int b64rmap_initialized = 0;
132 static uint8_t b64rmap[256];
133 
134 static const uint8_t b64rmap_special = 0xf0;
135 static const uint8_t b64rmap_end = 0xfd;
136 static const uint8_t b64rmap_space = 0xfe;
137 static const uint8_t b64rmap_invalid = 0xff;
138 
139 /**
140  * Initializing the reverse map is not thread safe.
141  * Which is fine for NSD. For now...
142  **/
143 static void
b64_initialize_rmap()144 b64_initialize_rmap ()
145 {
146 	int i;
147 
148 	/* Null: end of string, stop parsing */
149 	b64rmap[0] = b64rmap_end;
150 
151 	for (i = 1; i < 256; ++i) {
152 		/* Whitespaces */
153 		if (isspace(i))
154 			b64rmap[i] = b64rmap_space;
155 		/* Padding: stop parsing */
156 		else if (i == (unsigned char)Pad64)
157 			b64rmap[i] = b64rmap_end;
158 		/* Non-base64 char */
159 		else
160 			b64rmap[i] = b64rmap_invalid;
161 	}
162 
163 	/* Fill reverse mapping for base64 chars */
164 	for (i = 0; Base64[i] != '\0'; ++i)
165 		b64rmap[(uint8_t)Base64[i]] = i;
166 
167 	b64rmap_initialized = 1;
168 }
169 
170 static int
b64_pton_do(unsigned char const * src,uint8_t * target,size_t targsize)171 b64_pton_do(unsigned char const *src, uint8_t *target, size_t targsize)
172 {
173 	int tarindex, state, ch;
174 	uint8_t ofs;
175 
176 	state = 0;
177 	tarindex = 0;
178 
179 	while (1)
180 	{
181 		ch = *src++;
182 		ofs = b64rmap[ch];
183 
184 		if (ofs >= b64rmap_special) {
185 			/* Ignore whitespaces */
186 			if (ofs == b64rmap_space)
187 				continue;
188 			/* End of base64 characters */
189 			if (ofs == b64rmap_end)
190 				break;
191 			/* A non-base64 character. */
192 			return (-1);
193 		}
194 
195 		switch (state) {
196 		case 0:
197 			if ((size_t)tarindex >= targsize)
198 				return (-1);
199 			target[tarindex] = ofs << 2;
200 			state = 1;
201 			break;
202 		case 1:
203 			if ((size_t)tarindex + 1 >= targsize)
204 				return (-1);
205 			target[tarindex]   |=  ofs >> 4;
206 			target[tarindex+1]  = (ofs & 0x0f)
207 						<< 4 ;
208 			tarindex++;
209 			state = 2;
210 			break;
211 		case 2:
212 			if ((size_t)tarindex + 1 >= targsize)
213 				return (-1);
214 			target[tarindex]   |=  ofs >> 2;
215 			target[tarindex+1]  = (ofs & 0x03)
216 						<< 6;
217 			tarindex++;
218 			state = 3;
219 			break;
220 		case 3:
221 			if ((size_t)tarindex >= targsize)
222 				return (-1);
223 			target[tarindex] |= ofs;
224 			tarindex++;
225 			state = 0;
226 			break;
227 		default:
228 			abort();
229 		}
230 	}
231 
232 	/*
233 	 * We are done decoding Base-64 chars.  Let's see if we ended
234 	 * on a byte boundary, and/or with erroneous trailing characters.
235 	 */
236 
237 	if (ch == Pad64) {		/* We got a pad char. */
238 		ch = *src++;		/* Skip it, get next. */
239 		switch (state) {
240 		case 0:		/* Invalid = in first position */
241 		case 1:		/* Invalid = in second position */
242 			return (-1);
243 
244 		case 2:		/* Valid, means one byte of info */
245 			/* Skip any number of spaces. */
246 			for ((void)NULL; ch != '\0'; ch = *src++)
247 				if (b64rmap[ch] != b64rmap_space)
248 					break;
249 			/* Make sure there is another trailing = sign. */
250 			if (ch != Pad64)
251 				return (-1);
252 			ch = *src++;		/* Skip the = */
253 			/* Fall through to "single trailing =" case. */
254 			/* FALLTHROUGH */
255 
256 		case 3:		/* Valid, means two bytes of info */
257 			/*
258 			 * We know this char is an =.  Is there anything but
259 			 * whitespace after it?
260 			 */
261 			for ((void)NULL; ch != '\0'; ch = *src++)
262 				if (b64rmap[ch] != b64rmap_space)
263 					return (-1);
264 
265 			/*
266 			 * Now make sure for cases 2 and 3 that the "extra"
267 			 * bits that slopped past the last full byte were
268 			 * zeros.  If we don't check them, they become a
269 			 * subliminal channel.
270 			 */
271 			if (target[tarindex] != 0)
272 				return (-1);
273 		}
274 	} else {
275 		/*
276 		 * We ended by seeing the end of the string.  Make sure we
277 		 * have no partial bytes lying around.
278 		 */
279 		if (state != 0)
280 			return (-1);
281 	}
282 
283 	return (tarindex);
284 }
285 
286 
287 static int
b64_pton_len(unsigned char const * src)288 b64_pton_len(unsigned char const *src)
289 {
290 	int tarindex, state, ch;
291 	uint8_t ofs;
292 
293 	state = 0;
294 	tarindex = 0;
295 
296 	while (1)
297 	{
298 		ch = *src++;
299 		ofs = b64rmap[ch];
300 
301 		if (ofs >= b64rmap_special) {
302 			/* Ignore whitespaces */
303 			if (ofs == b64rmap_space)
304 				continue;
305 			/* End of base64 characters */
306 			if (ofs == b64rmap_end)
307 				break;
308 			/* A non-base64 character. */
309 			return (-1);
310 		}
311 
312 		switch (state) {
313 		case 0:
314 			state = 1;
315 			break;
316 		case 1:
317 			tarindex++;
318 			state = 2;
319 			break;
320 		case 2:
321 			tarindex++;
322 			state = 3;
323 			break;
324 		case 3:
325 			tarindex++;
326 			state = 0;
327 			break;
328 		default:
329 			abort();
330 		}
331 	}
332 
333 	/*
334 	 * We are done decoding Base-64 chars.  Let's see if we ended
335 	 * on a byte boundary, and/or with erroneous trailing characters.
336 	 */
337 
338 	if (ch == Pad64) {		/* We got a pad char. */
339 		ch = *src++;		/* Skip it, get next. */
340 		switch (state) {
341 		case 0:		/* Invalid = in first position */
342 		case 1:		/* Invalid = in second position */
343 			return (-1);
344 
345 		case 2:		/* Valid, means one byte of info */
346 			/* Skip any number of spaces. */
347 			for ((void)NULL; ch != '\0'; ch = *src++)
348 				if (b64rmap[ch] != b64rmap_space)
349 					break;
350 			/* Make sure there is another trailing = sign. */
351 			if (ch != Pad64)
352 				return (-1);
353 			ch = *src++;		/* Skip the = */
354 			/* Fall through to "single trailing =" case. */
355 			/* FALLTHROUGH */
356 
357 		case 3:		/* Valid, means two bytes of info */
358 			/*
359 			 * We know this char is an =.  Is there anything but
360 			 * whitespace after it?
361 			 */
362 			for ((void)NULL; ch != '\0'; ch = *src++)
363 				if (b64rmap[ch] != b64rmap_space)
364 					return (-1);
365 
366 		}
367 	} else {
368 		/*
369 		 * We ended by seeing the end of the string.  Make sure we
370 		 * have no partial bytes lying around.
371 		 */
372 		if (state != 0)
373 			return (-1);
374 	}
375 
376 	return (tarindex);
377 }
378 
379 
380 int
b64_pton(char const * src,uint8_t * target,size_t targsize)381 b64_pton(char const *src, uint8_t *target, size_t targsize)
382 {
383 	if (!b64rmap_initialized)
384 		b64_initialize_rmap ();
385 
386 	if (target)
387 		return b64_pton_do ((unsigned char const*)src, target, targsize);
388 	else
389 		return b64_pton_len ((unsigned char const*)src);
390 }
391