1*0Sstevel@tonic-gate /* crypto/asn1/a_utf8.c */
2*0Sstevel@tonic-gate /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
3*0Sstevel@tonic-gate * All rights reserved.
4*0Sstevel@tonic-gate *
5*0Sstevel@tonic-gate * This package is an SSL implementation written
6*0Sstevel@tonic-gate * by Eric Young (eay@cryptsoft.com).
7*0Sstevel@tonic-gate * The implementation was written so as to conform with Netscapes SSL.
8*0Sstevel@tonic-gate *
9*0Sstevel@tonic-gate * This library is free for commercial and non-commercial use as long as
10*0Sstevel@tonic-gate * the following conditions are aheared to. The following conditions
11*0Sstevel@tonic-gate * apply to all code found in this distribution, be it the RC4, RSA,
12*0Sstevel@tonic-gate * lhash, DES, etc., code; not just the SSL code. The SSL documentation
13*0Sstevel@tonic-gate * included with this distribution is covered by the same copyright terms
14*0Sstevel@tonic-gate * except that the holder is Tim Hudson (tjh@cryptsoft.com).
15*0Sstevel@tonic-gate *
16*0Sstevel@tonic-gate * Copyright remains Eric Young's, and as such any Copyright notices in
17*0Sstevel@tonic-gate * the code are not to be removed.
18*0Sstevel@tonic-gate * If this package is used in a product, Eric Young should be given attribution
19*0Sstevel@tonic-gate * as the author of the parts of the library used.
20*0Sstevel@tonic-gate * This can be in the form of a textual message at program startup or
21*0Sstevel@tonic-gate * in documentation (online or textual) provided with the package.
22*0Sstevel@tonic-gate *
23*0Sstevel@tonic-gate * Redistribution and use in source and binary forms, with or without
24*0Sstevel@tonic-gate * modification, are permitted provided that the following conditions
25*0Sstevel@tonic-gate * are met:
26*0Sstevel@tonic-gate * 1. Redistributions of source code must retain the copyright
27*0Sstevel@tonic-gate * notice, this list of conditions and the following disclaimer.
28*0Sstevel@tonic-gate * 2. Redistributions in binary form must reproduce the above copyright
29*0Sstevel@tonic-gate * notice, this list of conditions and the following disclaimer in the
30*0Sstevel@tonic-gate * documentation and/or other materials provided with the distribution.
31*0Sstevel@tonic-gate * 3. All advertising materials mentioning features or use of this software
32*0Sstevel@tonic-gate * must display the following acknowledgement:
33*0Sstevel@tonic-gate * "This product includes cryptographic software written by
34*0Sstevel@tonic-gate * Eric Young (eay@cryptsoft.com)"
35*0Sstevel@tonic-gate * The word 'cryptographic' can be left out if the rouines from the library
36*0Sstevel@tonic-gate * being used are not cryptographic related :-).
37*0Sstevel@tonic-gate * 4. If you include any Windows specific code (or a derivative thereof) from
38*0Sstevel@tonic-gate * the apps directory (application code) you must include an acknowledgement:
39*0Sstevel@tonic-gate * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
40*0Sstevel@tonic-gate *
41*0Sstevel@tonic-gate * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
42*0Sstevel@tonic-gate * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
43*0Sstevel@tonic-gate * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
44*0Sstevel@tonic-gate * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
45*0Sstevel@tonic-gate * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
46*0Sstevel@tonic-gate * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
47*0Sstevel@tonic-gate * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
48*0Sstevel@tonic-gate * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
49*0Sstevel@tonic-gate * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
50*0Sstevel@tonic-gate * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
51*0Sstevel@tonic-gate * SUCH DAMAGE.
52*0Sstevel@tonic-gate *
53*0Sstevel@tonic-gate * The licence and distribution terms for any publically available version or
54*0Sstevel@tonic-gate * derivative of this code cannot be changed. i.e. this code cannot simply be
55*0Sstevel@tonic-gate * copied and put under another distribution licence
56*0Sstevel@tonic-gate * [including the GNU Public Licence.]
57*0Sstevel@tonic-gate */
58*0Sstevel@tonic-gate
59*0Sstevel@tonic-gate #include <stdio.h>
60*0Sstevel@tonic-gate #include "cryptlib.h"
61*0Sstevel@tonic-gate #include <openssl/asn1.h>
62*0Sstevel@tonic-gate
63*0Sstevel@tonic-gate
64*0Sstevel@tonic-gate /* UTF8 utilities */
65*0Sstevel@tonic-gate
66*0Sstevel@tonic-gate /* This parses a UTF8 string one character at a time. It is passed a pointer
67*0Sstevel@tonic-gate * to the string and the length of the string. It sets 'value' to the value of
68*0Sstevel@tonic-gate * the current character. It returns the number of characters read or a
69*0Sstevel@tonic-gate * negative error code:
70*0Sstevel@tonic-gate * -1 = string too short
71*0Sstevel@tonic-gate * -2 = illegal character
72*0Sstevel@tonic-gate * -3 = subsequent characters not of the form 10xxxxxx
73*0Sstevel@tonic-gate * -4 = character encoded incorrectly (not minimal length).
74*0Sstevel@tonic-gate */
75*0Sstevel@tonic-gate
UTF8_getc(const unsigned char * str,int len,unsigned long * val)76*0Sstevel@tonic-gate int UTF8_getc(const unsigned char *str, int len, unsigned long *val)
77*0Sstevel@tonic-gate {
78*0Sstevel@tonic-gate const unsigned char *p;
79*0Sstevel@tonic-gate unsigned long value;
80*0Sstevel@tonic-gate int ret;
81*0Sstevel@tonic-gate if(len <= 0) return 0;
82*0Sstevel@tonic-gate p = str;
83*0Sstevel@tonic-gate
84*0Sstevel@tonic-gate /* Check syntax and work out the encoded value (if correct) */
85*0Sstevel@tonic-gate if((*p & 0x80) == 0) {
86*0Sstevel@tonic-gate value = *p++ & 0x7f;
87*0Sstevel@tonic-gate ret = 1;
88*0Sstevel@tonic-gate } else if((*p & 0xe0) == 0xc0) {
89*0Sstevel@tonic-gate if(len < 2) return -1;
90*0Sstevel@tonic-gate if((p[1] & 0xc0) != 0x80) return -3;
91*0Sstevel@tonic-gate value = (*p++ & 0x1f) << 6;
92*0Sstevel@tonic-gate value |= *p++ & 0x3f;
93*0Sstevel@tonic-gate if(value < 0x80) return -4;
94*0Sstevel@tonic-gate ret = 2;
95*0Sstevel@tonic-gate } else if((*p & 0xf0) == 0xe0) {
96*0Sstevel@tonic-gate if(len < 3) return -1;
97*0Sstevel@tonic-gate if( ((p[1] & 0xc0) != 0x80)
98*0Sstevel@tonic-gate || ((p[2] & 0xc0) != 0x80) ) return -3;
99*0Sstevel@tonic-gate value = (*p++ & 0xf) << 12;
100*0Sstevel@tonic-gate value |= (*p++ & 0x3f) << 6;
101*0Sstevel@tonic-gate value |= *p++ & 0x3f;
102*0Sstevel@tonic-gate if(value < 0x800) return -4;
103*0Sstevel@tonic-gate ret = 3;
104*0Sstevel@tonic-gate } else if((*p & 0xf8) == 0xf0) {
105*0Sstevel@tonic-gate if(len < 4) return -1;
106*0Sstevel@tonic-gate if( ((p[1] & 0xc0) != 0x80)
107*0Sstevel@tonic-gate || ((p[2] & 0xc0) != 0x80)
108*0Sstevel@tonic-gate || ((p[3] & 0xc0) != 0x80) ) return -3;
109*0Sstevel@tonic-gate value = ((unsigned long)(*p++ & 0x7)) << 18;
110*0Sstevel@tonic-gate value |= (*p++ & 0x3f) << 12;
111*0Sstevel@tonic-gate value |= (*p++ & 0x3f) << 6;
112*0Sstevel@tonic-gate value |= *p++ & 0x3f;
113*0Sstevel@tonic-gate if(value < 0x10000) return -4;
114*0Sstevel@tonic-gate ret = 4;
115*0Sstevel@tonic-gate } else if((*p & 0xfc) == 0xf8) {
116*0Sstevel@tonic-gate if(len < 5) return -1;
117*0Sstevel@tonic-gate if( ((p[1] & 0xc0) != 0x80)
118*0Sstevel@tonic-gate || ((p[2] & 0xc0) != 0x80)
119*0Sstevel@tonic-gate || ((p[3] & 0xc0) != 0x80)
120*0Sstevel@tonic-gate || ((p[4] & 0xc0) != 0x80) ) return -3;
121*0Sstevel@tonic-gate value = ((unsigned long)(*p++ & 0x3)) << 24;
122*0Sstevel@tonic-gate value |= ((unsigned long)(*p++ & 0x3f)) << 18;
123*0Sstevel@tonic-gate value |= ((unsigned long)(*p++ & 0x3f)) << 12;
124*0Sstevel@tonic-gate value |= (*p++ & 0x3f) << 6;
125*0Sstevel@tonic-gate value |= *p++ & 0x3f;
126*0Sstevel@tonic-gate if(value < 0x200000) return -4;
127*0Sstevel@tonic-gate ret = 5;
128*0Sstevel@tonic-gate } else if((*p & 0xfe) == 0xfc) {
129*0Sstevel@tonic-gate if(len < 6) return -1;
130*0Sstevel@tonic-gate if( ((p[1] & 0xc0) != 0x80)
131*0Sstevel@tonic-gate || ((p[2] & 0xc0) != 0x80)
132*0Sstevel@tonic-gate || ((p[3] & 0xc0) != 0x80)
133*0Sstevel@tonic-gate || ((p[4] & 0xc0) != 0x80)
134*0Sstevel@tonic-gate || ((p[5] & 0xc0) != 0x80) ) return -3;
135*0Sstevel@tonic-gate value = ((unsigned long)(*p++ & 0x1)) << 30;
136*0Sstevel@tonic-gate value |= ((unsigned long)(*p++ & 0x3f)) << 24;
137*0Sstevel@tonic-gate value |= ((unsigned long)(*p++ & 0x3f)) << 18;
138*0Sstevel@tonic-gate value |= ((unsigned long)(*p++ & 0x3f)) << 12;
139*0Sstevel@tonic-gate value |= (*p++ & 0x3f) << 6;
140*0Sstevel@tonic-gate value |= *p++ & 0x3f;
141*0Sstevel@tonic-gate if(value < 0x4000000) return -4;
142*0Sstevel@tonic-gate ret = 6;
143*0Sstevel@tonic-gate } else return -2;
144*0Sstevel@tonic-gate *val = value;
145*0Sstevel@tonic-gate return ret;
146*0Sstevel@tonic-gate }
147*0Sstevel@tonic-gate
148*0Sstevel@tonic-gate /* This takes a character 'value' and writes the UTF8 encoded value in
149*0Sstevel@tonic-gate * 'str' where 'str' is a buffer containing 'len' characters. Returns
150*0Sstevel@tonic-gate * the number of characters written or -1 if 'len' is too small. 'str' can
151*0Sstevel@tonic-gate * be set to NULL in which case it just returns the number of characters.
152*0Sstevel@tonic-gate * It will need at most 6 characters.
153*0Sstevel@tonic-gate */
154*0Sstevel@tonic-gate
UTF8_putc(unsigned char * str,int len,unsigned long value)155*0Sstevel@tonic-gate int UTF8_putc(unsigned char *str, int len, unsigned long value)
156*0Sstevel@tonic-gate {
157*0Sstevel@tonic-gate if(!str) len = 6; /* Maximum we will need */
158*0Sstevel@tonic-gate else if(len <= 0) return -1;
159*0Sstevel@tonic-gate if(value < 0x80) {
160*0Sstevel@tonic-gate if(str) *str = (unsigned char)value;
161*0Sstevel@tonic-gate return 1;
162*0Sstevel@tonic-gate }
163*0Sstevel@tonic-gate if(value < 0x800) {
164*0Sstevel@tonic-gate if(len < 2) return -1;
165*0Sstevel@tonic-gate if(str) {
166*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 6) & 0x1f) | 0xc0);
167*0Sstevel@tonic-gate *str = (unsigned char)((value & 0x3f) | 0x80);
168*0Sstevel@tonic-gate }
169*0Sstevel@tonic-gate return 2;
170*0Sstevel@tonic-gate }
171*0Sstevel@tonic-gate if(value < 0x10000) {
172*0Sstevel@tonic-gate if(len < 3) return -1;
173*0Sstevel@tonic-gate if(str) {
174*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 12) & 0xf) | 0xe0);
175*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 6) & 0x3f) | 0x80);
176*0Sstevel@tonic-gate *str = (unsigned char)((value & 0x3f) | 0x80);
177*0Sstevel@tonic-gate }
178*0Sstevel@tonic-gate return 3;
179*0Sstevel@tonic-gate }
180*0Sstevel@tonic-gate if(value < 0x200000) {
181*0Sstevel@tonic-gate if(len < 4) return -1;
182*0Sstevel@tonic-gate if(str) {
183*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 18) & 0x7) | 0xf0);
184*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 12) & 0x3f) | 0x80);
185*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 6) & 0x3f) | 0x80);
186*0Sstevel@tonic-gate *str = (unsigned char)((value & 0x3f) | 0x80);
187*0Sstevel@tonic-gate }
188*0Sstevel@tonic-gate return 4;
189*0Sstevel@tonic-gate }
190*0Sstevel@tonic-gate if(value < 0x4000000) {
191*0Sstevel@tonic-gate if(len < 5) return -1;
192*0Sstevel@tonic-gate if(str) {
193*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 24) & 0x3) | 0xf8);
194*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 18) & 0x3f) | 0x80);
195*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 12) & 0x3f) | 0x80);
196*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 6) & 0x3f) | 0x80);
197*0Sstevel@tonic-gate *str = (unsigned char)((value & 0x3f) | 0x80);
198*0Sstevel@tonic-gate }
199*0Sstevel@tonic-gate return 5;
200*0Sstevel@tonic-gate }
201*0Sstevel@tonic-gate if(len < 6) return -1;
202*0Sstevel@tonic-gate if(str) {
203*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 30) & 0x1) | 0xfc);
204*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 24) & 0x3f) | 0x80);
205*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 18) & 0x3f) | 0x80);
206*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 12) & 0x3f) | 0x80);
207*0Sstevel@tonic-gate *str++ = (unsigned char)(((value >> 6) & 0x3f) | 0x80);
208*0Sstevel@tonic-gate *str = (unsigned char)((value & 0x3f) | 0x80);
209*0Sstevel@tonic-gate }
210*0Sstevel@tonic-gate return 6;
211*0Sstevel@tonic-gate }
212