xref: /openssl/crypto/asn1/a_mbstr.c (revision 73ebaac8)
1 /*
2  * Copyright 1999-2021 The OpenSSL Project Authors. All Rights Reserved.
3  *
4  * Licensed under the Apache License 2.0 (the "License").  You may not use
5  * this file except in compliance with the License.  You can obtain a copy
6  * in the file LICENSE in the source distribution or at
7  * https://www.openssl.org/source/license.html
8  */
9 
10 #include <stdio.h>
11 #include "crypto/ctype.h"
12 #include "internal/cryptlib.h"
13 #include "internal/unicode.h"
14 #include <openssl/asn1.h>
15 
16 static int traverse_string(const unsigned char *p, int len, int inform,
17                            int (*rfunc) (unsigned long value, void *in),
18                            void *arg);
19 static int in_utf8(unsigned long value, void *arg);
20 static int out_utf8(unsigned long value, void *arg);
21 static int type_str(unsigned long value, void *arg);
22 static int cpy_asc(unsigned long value, void *arg);
23 static int cpy_bmp(unsigned long value, void *arg);
24 static int cpy_univ(unsigned long value, void *arg);
25 static int cpy_utf8(unsigned long value, void *arg);
26 
27 /*
28  * These functions take a string in UTF8, ASCII or multibyte form and a mask
29  * of permissible ASN1 string types. It then works out the minimal type
30  * (using the order Numeric < Printable < IA5 < T61 < BMP < Universal < UTF8)
31  * and creates a string of the correct type with the supplied data. Yes this is
32  * horrible: it has to be :-( The 'ncopy' form checks minimum and maximum
33  * size limits too.
34  */
35 
ASN1_mbstring_copy(ASN1_STRING ** out,const unsigned char * in,int len,int inform,unsigned long mask)36 int ASN1_mbstring_copy(ASN1_STRING **out, const unsigned char *in, int len,
37                        int inform, unsigned long mask)
38 {
39     return ASN1_mbstring_ncopy(out, in, len, inform, mask, 0, 0);
40 }
41 
ASN1_mbstring_ncopy(ASN1_STRING ** out,const unsigned char * in,int len,int inform,unsigned long mask,long minsize,long maxsize)42 int ASN1_mbstring_ncopy(ASN1_STRING **out, const unsigned char *in, int len,
43                         int inform, unsigned long mask,
44                         long minsize, long maxsize)
45 {
46     int str_type;
47     int ret;
48     char free_out;
49     int outform, outlen = 0;
50     ASN1_STRING *dest;
51     unsigned char *p;
52     int nchar;
53     int (*cpyfunc) (unsigned long, void *) = NULL;
54     if (len == -1)
55         len = strlen((const char *)in);
56     if (!mask)
57         mask = DIRSTRING_TYPE;
58     if (len < 0)
59         return -1;
60 
61     /* First do a string check and work out the number of characters */
62     switch (inform) {
63 
64     case MBSTRING_BMP:
65         if (len & 1) {
66             ERR_raise(ERR_LIB_ASN1, ASN1_R_INVALID_BMPSTRING_LENGTH);
67             return -1;
68         }
69         nchar = len >> 1;
70         break;
71 
72     case MBSTRING_UNIV:
73         if (len & 3) {
74             ERR_raise(ERR_LIB_ASN1, ASN1_R_INVALID_UNIVERSALSTRING_LENGTH);
75             return -1;
76         }
77         nchar = len >> 2;
78         break;
79 
80     case MBSTRING_UTF8:
81         nchar = 0;
82         /* This counts the characters and does utf8 syntax checking */
83         ret = traverse_string(in, len, MBSTRING_UTF8, in_utf8, &nchar);
84         if (ret < 0) {
85             ERR_raise(ERR_LIB_ASN1, ASN1_R_INVALID_UTF8STRING);
86             return -1;
87         }
88         break;
89 
90     case MBSTRING_ASC:
91         nchar = len;
92         break;
93 
94     default:
95         ERR_raise(ERR_LIB_ASN1, ASN1_R_UNKNOWN_FORMAT);
96         return -1;
97     }
98 
99     if ((minsize > 0) && (nchar < minsize)) {
100         ERR_raise_data(ERR_LIB_ASN1, ASN1_R_STRING_TOO_SHORT,
101                        "minsize=%ld", minsize);
102         return -1;
103     }
104 
105     if ((maxsize > 0) && (nchar > maxsize)) {
106         ERR_raise_data(ERR_LIB_ASN1, ASN1_R_STRING_TOO_LONG,
107                        "maxsize=%ld", maxsize);
108         return -1;
109     }
110 
111     /* Now work out minimal type (if any) */
112     if (traverse_string(in, len, inform, type_str, &mask) < 0) {
113         ERR_raise(ERR_LIB_ASN1, ASN1_R_ILLEGAL_CHARACTERS);
114         return -1;
115     }
116 
117     /* Now work out output format and string type */
118     outform = MBSTRING_ASC;
119     if (mask & B_ASN1_NUMERICSTRING)
120         str_type = V_ASN1_NUMERICSTRING;
121     else if (mask & B_ASN1_PRINTABLESTRING)
122         str_type = V_ASN1_PRINTABLESTRING;
123     else if (mask & B_ASN1_IA5STRING)
124         str_type = V_ASN1_IA5STRING;
125     else if (mask & B_ASN1_T61STRING)
126         str_type = V_ASN1_T61STRING;
127     else if (mask & B_ASN1_BMPSTRING) {
128         str_type = V_ASN1_BMPSTRING;
129         outform = MBSTRING_BMP;
130     } else if (mask & B_ASN1_UNIVERSALSTRING) {
131         str_type = V_ASN1_UNIVERSALSTRING;
132         outform = MBSTRING_UNIV;
133     } else {
134         str_type = V_ASN1_UTF8STRING;
135         outform = MBSTRING_UTF8;
136     }
137     if (!out)
138         return str_type;
139     if (*out) {
140         free_out = 0;
141         dest = *out;
142         ASN1_STRING_set0(dest, NULL, 0);
143         dest->type = str_type;
144     } else {
145         free_out = 1;
146         dest = ASN1_STRING_type_new(str_type);
147         if (dest == NULL) {
148             ERR_raise(ERR_LIB_ASN1, ERR_R_ASN1_LIB);
149             return -1;
150         }
151         *out = dest;
152     }
153     /* If both the same type just copy across */
154     if (inform == outform) {
155         if (!ASN1_STRING_set(dest, in, len)) {
156             if (free_out) {
157                 ASN1_STRING_free(dest);
158                 *out = NULL;
159             }
160             ERR_raise(ERR_LIB_ASN1, ERR_R_ASN1_LIB);
161             return -1;
162         }
163         return str_type;
164     }
165 
166     /* Work out how much space the destination will need */
167     switch (outform) {
168     case MBSTRING_ASC:
169         outlen = nchar;
170         cpyfunc = cpy_asc;
171         break;
172 
173     case MBSTRING_BMP:
174         outlen = nchar << 1;
175         cpyfunc = cpy_bmp;
176         break;
177 
178     case MBSTRING_UNIV:
179         outlen = nchar << 2;
180         cpyfunc = cpy_univ;
181         break;
182 
183     case MBSTRING_UTF8:
184         outlen = 0;
185         traverse_string(in, len, inform, out_utf8, &outlen);
186         cpyfunc = cpy_utf8;
187         break;
188     }
189     if ((p = OPENSSL_malloc(outlen + 1)) == NULL) {
190         if (free_out) {
191             ASN1_STRING_free(dest);
192             *out = NULL;
193         }
194         return -1;
195     }
196     dest->length = outlen;
197     dest->data = p;
198     p[outlen] = 0;
199     traverse_string(in, len, inform, cpyfunc, &p);
200     return str_type;
201 }
202 
203 /*
204  * This function traverses a string and passes the value of each character to
205  * an optional function along with a void * argument.
206  */
207 
traverse_string(const unsigned char * p,int len,int inform,int (* rfunc)(unsigned long value,void * in),void * arg)208 static int traverse_string(const unsigned char *p, int len, int inform,
209                            int (*rfunc) (unsigned long value, void *in),
210                            void *arg)
211 {
212     unsigned long value;
213     int ret;
214     while (len) {
215         if (inform == MBSTRING_ASC) {
216             value = *p++;
217             len--;
218         } else if (inform == MBSTRING_BMP) {
219             value = *p++ << 8;
220             value |= *p++;
221             len -= 2;
222         } else if (inform == MBSTRING_UNIV) {
223             value = ((unsigned long)*p++) << 24;
224             value |= ((unsigned long)*p++) << 16;
225             value |= *p++ << 8;
226             value |= *p++;
227             len -= 4;
228         } else {
229             ret = UTF8_getc(p, len, &value);
230             if (ret < 0)
231                 return -1;
232             len -= ret;
233             p += ret;
234         }
235         if (rfunc) {
236             ret = rfunc(value, arg);
237             if (ret <= 0)
238                 return ret;
239         }
240     }
241     return 1;
242 }
243 
244 /* Various utility functions for traverse_string */
245 
246 /* Just count number of characters */
247 
in_utf8(unsigned long value,void * arg)248 static int in_utf8(unsigned long value, void *arg)
249 {
250     int *nchar;
251 
252     if (!is_unicode_valid(value))
253         return -2;
254     nchar = arg;
255     (*nchar)++;
256     return 1;
257 }
258 
259 /* Determine size of output as a UTF8 String */
260 
out_utf8(unsigned long value,void * arg)261 static int out_utf8(unsigned long value, void *arg)
262 {
263     int *outlen, len;
264 
265     len = UTF8_putc(NULL, -1, value);
266     if (len <= 0)
267         return len;
268     outlen = arg;
269     *outlen += len;
270     return 1;
271 }
272 
273 /*
274  * Determine the "type" of a string: check each character against a supplied
275  * "mask".
276  */
277 
type_str(unsigned long value,void * arg)278 static int type_str(unsigned long value, void *arg)
279 {
280     unsigned long types = *((unsigned long *)arg);
281     const int native = value > INT_MAX ? INT_MAX : ossl_fromascii(value);
282 
283     if ((types & B_ASN1_NUMERICSTRING) && !(ossl_isdigit(native)
284                                             || native == ' '))
285         types &= ~B_ASN1_NUMERICSTRING;
286     if ((types & B_ASN1_PRINTABLESTRING) && !ossl_isasn1print(native))
287         types &= ~B_ASN1_PRINTABLESTRING;
288     if ((types & B_ASN1_IA5STRING) && !ossl_isascii(native))
289         types &= ~B_ASN1_IA5STRING;
290     if ((types & B_ASN1_T61STRING) && (value > 0xff))
291         types &= ~B_ASN1_T61STRING;
292     if ((types & B_ASN1_BMPSTRING) && (value > 0xffff))
293         types &= ~B_ASN1_BMPSTRING;
294     if ((types & B_ASN1_UTF8STRING) && !is_unicode_valid(value))
295         types &= ~B_ASN1_UTF8STRING;
296     if (!types)
297         return -1;
298     *((unsigned long *)arg) = types;
299     return 1;
300 }
301 
302 /* Copy one byte per character ASCII like strings */
303 
cpy_asc(unsigned long value,void * arg)304 static int cpy_asc(unsigned long value, void *arg)
305 {
306     unsigned char **p, *q;
307     p = arg;
308     q = *p;
309     *q = (unsigned char)value;
310     (*p)++;
311     return 1;
312 }
313 
314 /* Copy two byte per character BMPStrings */
315 
cpy_bmp(unsigned long value,void * arg)316 static int cpy_bmp(unsigned long value, void *arg)
317 {
318     unsigned char **p, *q;
319     p = arg;
320     q = *p;
321     *q++ = (unsigned char)((value >> 8) & 0xff);
322     *q = (unsigned char)(value & 0xff);
323     *p += 2;
324     return 1;
325 }
326 
327 /* Copy four byte per character UniversalStrings */
328 
cpy_univ(unsigned long value,void * arg)329 static int cpy_univ(unsigned long value, void *arg)
330 {
331     unsigned char **p, *q;
332     p = arg;
333     q = *p;
334     *q++ = (unsigned char)((value >> 24) & 0xff);
335     *q++ = (unsigned char)((value >> 16) & 0xff);
336     *q++ = (unsigned char)((value >> 8) & 0xff);
337     *q = (unsigned char)(value & 0xff);
338     *p += 4;
339     return 1;
340 }
341 
342 /* Copy to a UTF8String */
343 
cpy_utf8(unsigned long value,void * arg)344 static int cpy_utf8(unsigned long value, void *arg)
345 {
346     unsigned char **p;
347     int ret;
348     p = arg;
349     /* We already know there is enough room so pass 0xff as the length */
350     ret = UTF8_putc(*p, 0xff, value);
351     *p += ret;
352     return 1;
353 }
354