crypto/asn1/a_mbstr.c

   1 /* a_mbstr.c */
   2 /*
   3  * Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL project
   4  * 1999.
   5  */
   6 /* ====================================================================
   7  * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
   8  *
   9  * Redistribution and use in source and binary forms, with or without
  10  * modification, are permitted provided that the following conditions
  11  * are met:
  12  *
  13  * 1. Redistributions of source code must retain the above copyright
  14  *    notice, this list of conditions and the following disclaimer.
  15  *
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in
  18  *    the documentation and/or other materials provided with the
  19  *    distribution.
  20  *
  21  * 3. All advertising materials mentioning features or use of this
  22  *    software must display the following acknowledgment:
  23  *    "This product includes software developed by the OpenSSL Project
  24  *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
  25  *
  26  * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
  27  *    endorse or promote products derived from this software without
  28  *    prior written permission. For written permission, please contact
  29  *    licensing@OpenSSL.org.
  30  *
  31  * 5. Products derived from this software may not be called "OpenSSL"
  32  *    nor may "OpenSSL" appear in their names without prior written
  33  *    permission of the OpenSSL Project.
  34  *
  35  * 6. Redistributions of any form whatsoever must retain the following
  36  *    acknowledgment:
  37  *    "This product includes software developed by the OpenSSL Project
  38  *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
  39  *
  40  * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
  41  * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  42  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
  43  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
  44  * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  45  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
  46  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  47  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  48  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
  49  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  50  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
  51  * OF THE POSSIBILITY OF SUCH DAMAGE.
  52  * ====================================================================
  53  *
  54  * This product includes cryptographic software written by Eric Young
  55  * (eay@cryptsoft.com).  This product includes software written by Tim
  56  * Hudson (tjh@cryptsoft.com).
  57  *
  58  */
  59
  60 #include <stdio.h>
  61 #include <ctype.h>
  62 #include "internal/cryptlib.h"
  63 #include <openssl/asn1.h>
  64
  65 static int traverse_string(const unsigned char *p, int len, int inform,
  66                            int (*rfunc) (unsigned long value, void *in),
  67                            void *arg);
  68 static int in_utf8(unsigned long value, void *arg);
  69 static int out_utf8(unsigned long value, void *arg);
  70 static int type_str(unsigned long value, void *arg);
  71 static int cpy_asc(unsigned long value, void *arg);
  72 static int cpy_bmp(unsigned long value, void *arg);
  73 static int cpy_univ(unsigned long value, void *arg);
  74 static int cpy_utf8(unsigned long value, void *arg);
  75 static int is_numeric(unsigned long value);
  76 static int is_printable(unsigned long value);
  77
  78 /*
  79  * These functions take a string in UTF8, ASCII or multibyte form and a mask
  80  * of permissible ASN1 string types. It then works out the minimal type
  81  * (using the order Numeric < Printable < IA5 < T61 < BMP < Universal < UTF8)
  82  * and creates a string of the correct type with the supplied data. Yes this is
  83  * horrible: it has to be :-( The 'ncopy' form checks minimum and maximum
  84  * size limits too.
  85  */
  86
  87 int ASN1_mbstring_copy(ASN1_STRING **out, const unsigned char *in, int len,
  88                        int inform, unsigned long mask)
  89 {
  90     return ASN1_mbstring_ncopy(out, in, len, inform, mask, 0, 0);
  91 }
  92
  93 int ASN1_mbstring_ncopy(ASN1_STRING **out, const unsigned char *in, int len,
  94                         int inform, unsigned long mask,
  95                         long minsize, long maxsize)
  96 {
  97     int str_type;
  98     int ret;
  99     char free_out;
 100     int outform, outlen = 0;
 101     ASN1_STRING *dest;
 102     unsigned char *p;
 103     int nchar;
 104     char strbuf[32];
 105     int (*cpyfunc) (unsigned long, void *) = NULL;
 106     if (len == -1)
 107         len = strlen((const char *)in);
 108     if (!mask)
 109         mask = DIRSTRING_TYPE;
 110
 111     /* First do a string check and work out the number of characters */
 112     switch (inform) {
 113
 114     case MBSTRING_BMP:
 115         if (len & 1) {
 116             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY,
 117                     ASN1_R_INVALID_BMPSTRING_LENGTH);
 118             return -1;
 119         }
 120         nchar = len >> 1;
 121         break;
 122
 123     case MBSTRING_UNIV:
 124         if (len & 3) {
 125             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY,
 126                     ASN1_R_INVALID_UNIVERSALSTRING_LENGTH);
 127             return -1;
 128         }
 129         nchar = len >> 2;
 130         break;
 131
 132     case MBSTRING_UTF8:
 133         nchar = 0;
 134         /* This counts the characters and does utf8 syntax checking */
 135         ret = traverse_string(in, len, MBSTRING_UTF8, in_utf8, &nchar);
 136         if (ret < 0) {
 137             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_INVALID_UTF8STRING);
 138             return -1;
 139         }
 140         break;
 141
 142     case MBSTRING_ASC:
 143         nchar = len;
 144         break;
 145
 146     default:
 147         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_UNKNOWN_FORMAT);
 148         return -1;
 149     }
 150
 151     if ((minsize > 0) && (nchar < minsize)) {
 152         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_STRING_TOO_SHORT);
 153         BIO_snprintf(strbuf, sizeof strbuf, "%ld", minsize);
 154         ERR_add_error_data(2, "minsize=", strbuf);
 155         return -1;
 156     }
 157
 158     if ((maxsize > 0) && (nchar > maxsize)) {
 159         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_STRING_TOO_LONG);
 160         BIO_snprintf(strbuf, sizeof strbuf, "%ld", maxsize);
 161         ERR_add_error_data(2, "maxsize=", strbuf);
 162         return -1;
 163     }
 164
 165     /* Now work out minimal type (if any) */
 166     if (traverse_string(in, len, inform, type_str, &mask) < 0) {
 167         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_ILLEGAL_CHARACTERS);
 168         return -1;
 169     }
 170
 171     /* Now work out output format and string type */
 172     outform = MBSTRING_ASC;
 173     if (mask & B_ASN1_NUMERICSTRING)
 174         str_type = V_ASN1_NUMERICSTRING;
 175     else if (mask & B_ASN1_PRINTABLESTRING)
 176         str_type = V_ASN1_PRINTABLESTRING;
 177     else if (mask & B_ASN1_IA5STRING)
 178         str_type = V_ASN1_IA5STRING;
 179     else if (mask & B_ASN1_T61STRING)
 180         str_type = V_ASN1_T61STRING;
 181     else if (mask & B_ASN1_BMPSTRING) {
 182         str_type = V_ASN1_BMPSTRING;
 183         outform = MBSTRING_BMP;
 184     } else if (mask & B_ASN1_UNIVERSALSTRING) {
 185         str_type = V_ASN1_UNIVERSALSTRING;
 186         outform = MBSTRING_UNIV;
 187     } else {
 188         str_type = V_ASN1_UTF8STRING;
 189         outform = MBSTRING_UTF8;
 190     }
 191     if (!out)
 192         return str_type;
 193     if (*out) {
 194         free_out = 0;
 195         dest = *out;
 196         OPENSSL_free(dest->data);
 197         dest->data = NULL;
 198         dest->length = 0;
 199         dest->type = str_type;
 200     } else {
 201         free_out = 1;
 202         dest = ASN1_STRING_type_new(str_type);
 203         if (!dest) {
 204             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 205             return -1;
 206         }
 207         *out = dest;
 208     }
 209     /* If both the same type just copy across */
 210     if (inform == outform) {
 211         if (!ASN1_STRING_set(dest, in, len)) {
 212             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 213             return -1;
 214         }
 215         return str_type;
 216     }
 217
 218     /* Work out how much space the destination will need */
 219     switch (outform) {
 220     case MBSTRING_ASC:
 221         outlen = nchar;
 222         cpyfunc = cpy_asc;
 223         break;
 224
 225     case MBSTRING_BMP:
 226         outlen = nchar << 1;
 227         cpyfunc = cpy_bmp;
 228         break;
 229
 230     case MBSTRING_UNIV:
 231         outlen = nchar << 2;
 232         cpyfunc = cpy_univ;
 233         break;
 234
 235     case MBSTRING_UTF8:
 236         outlen = 0;
 237         traverse_string(in, len, inform, out_utf8, &outlen);
 238         cpyfunc = cpy_utf8;
 239         break;
 240     }
 241     if ((p = OPENSSL_malloc(outlen + 1)) == NULL) {
 242         if (free_out)
 243             ASN1_STRING_free(dest);
 244         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 245         return -1;
 246     }
 247     dest->length = outlen;
 248     dest->data = p;
 249     p[outlen] = 0;
 250     traverse_string(in, len, inform, cpyfunc, &p);
 251     return str_type;
 252 }
 253
 254 /*
 255  * This function traverses a string and passes the value of each character to
 256  * an optional function along with a void * argument.
 257  */
 258
 259 static int traverse_string(const unsigned char *p, int len, int inform,
 260                            int (*rfunc) (unsigned long value, void *in),
 261                            void *arg)
 262 {
 263     unsigned long value;
 264     int ret;
 265     while (len) {
 266         if (inform == MBSTRING_ASC) {
 267             value = *p++;
 268             len--;
 269         } else if (inform == MBSTRING_BMP) {
 270             value = *p++ << 8;
 271             value |= *p++;
 272             len -= 2;
 273         } else if (inform == MBSTRING_UNIV) {
 274             value = ((unsigned long)*p++) << 24;
 275             value |= ((unsigned long)*p++) << 16;
 276             value |= *p++ << 8;
 277             value |= *p++;
 278             len -= 4;
 279         } else {
 280             ret = UTF8_getc(p, len, &value);
 281             if (ret < 0)
 282                 return -1;
 283             len -= ret;
 284             p += ret;
 285         }
 286         if (rfunc) {
 287             ret = rfunc(value, arg);
 288             if (ret <= 0)
 289                 return ret;
 290         }
 291     }
 292     return 1;
 293 }
 294
 295 /* Various utility functions for traverse_string */
 296
 297 /* Just count number of characters */
 298
 299 static int in_utf8(unsigned long value, void *arg)
 300 {
 301     int *nchar;
 302     nchar = arg;
 303     (*nchar)++;
 304     return 1;
 305 }
 306
 307 /* Determine size of output as a UTF8 String */
 308
 309 static int out_utf8(unsigned long value, void *arg)
 310 {
 311     int *outlen;
 312     outlen = arg;
 313     *outlen += UTF8_putc(NULL, -1, value);
 314     return 1;
 315 }
 316
 317 /*
 318  * Determine the "type" of a string: check each character against a supplied
 319  * "mask".
 320  */
 321
 322 static int type_str(unsigned long value, void *arg)
 323 {
 324     unsigned long types;
 325     types = *((unsigned long *)arg);
 326     if ((types & B_ASN1_NUMERICSTRING) && !is_numeric(value))
 327         types &= ~B_ASN1_NUMERICSTRING;
 328     if ((types & B_ASN1_PRINTABLESTRING) && !is_printable(value))
 329         types &= ~B_ASN1_PRINTABLESTRING;
 330     if ((types & B_ASN1_IA5STRING) && (value > 127))
 331         types &= ~B_ASN1_IA5STRING;
 332     if ((types & B_ASN1_T61STRING) && (value > 0xff))
 333         types &= ~B_ASN1_T61STRING;
 334     if ((types & B_ASN1_BMPSTRING) && (value > 0xffff))
 335         types &= ~B_ASN1_BMPSTRING;
 336     if (!types)
 337         return -1;
 338     *((unsigned long *)arg) = types;
 339     return 1;
 340 }
 341
 342 /* Copy one byte per character ASCII like strings */
 343
 344 static int cpy_asc(unsigned long value, void *arg)
 345 {
 346     unsigned char **p, *q;
 347     p = arg;
 348     q = *p;
 349     *q = (unsigned char)value;
 350     (*p)++;
 351     return 1;
 352 }
 353
 354 /* Copy two byte per character BMPStrings */
 355
 356 static int cpy_bmp(unsigned long value, void *arg)
 357 {
 358     unsigned char **p, *q;
 359     p = arg;
 360     q = *p;
 361     *q++ = (unsigned char)((value >> 8) & 0xff);
 362     *q = (unsigned char)(value & 0xff);
 363     *p += 2;
 364     return 1;
 365 }
 366
 367 /* Copy four byte per character UniversalStrings */
 368
 369 static int cpy_univ(unsigned long value, void *arg)
 370 {
 371     unsigned char **p, *q;
 372     p = arg;
 373     q = *p;
 374     *q++ = (unsigned char)((value >> 24) & 0xff);
 375     *q++ = (unsigned char)((value >> 16) & 0xff);
 376     *q++ = (unsigned char)((value >> 8) & 0xff);
 377     *q = (unsigned char)(value & 0xff);
 378     *p += 4;
 379     return 1;
 380 }
 381
 382 /* Copy to a UTF8String */
 383
 384 static int cpy_utf8(unsigned long value, void *arg)
 385 {
 386     unsigned char **p;
 387     int ret;
 388     p = arg;
 389     /* We already know there is enough room so pass 0xff as the length */
 390     ret = UTF8_putc(*p, 0xff, value);
 391     *p += ret;
 392     return 1;
 393 }
 394
 395 /* Return 1 if the character is permitted in a PrintableString */
 396 static int is_printable(unsigned long value)
 397 {
 398     int ch;
 399     if (value > 0x7f)
 400         return 0;
 401     ch = (int)value;
 402     /*
 403      * Note: we can't use 'isalnum' because certain accented characters may
 404      * count as alphanumeric in some environments.
 405      */
 406 #ifndef CHARSET_EBCDIC
 407     if ((ch >= 'a') && (ch <= 'z'))
 408         return 1;
 409     if ((ch >= 'A') && (ch <= 'Z'))
 410         return 1;
 411     if ((ch >= '0') && (ch <= '9'))
 412         return 1;
 413     if ((ch == ' ') || strchr("'()+,-./:=?", ch))
 414         return 1;
 415 #else                           /* CHARSET_EBCDIC */
 416     if ((ch >= os_toascii['a']) && (ch <= os_toascii['z']))
 417         return 1;
 418     if ((ch >= os_toascii['A']) && (ch <= os_toascii['Z']))
 419         return 1;
 420     if ((ch >= os_toascii['0']) && (ch <= os_toascii['9']))
 421         return 1;
 422     if ((ch == os_toascii[' ']) || strchr("'()+,-./:=?", os_toebcdic[ch]))
 423         return 1;
 424 #endif                          /* CHARSET_EBCDIC */
 425     return 0;
 426 }
 427
 428 /* Return 1 if the character is a digit or space */
 429 static int is_numeric(unsigned long value)
 430 {
 431     int ch;
 432     if (value > 0x7f)
 433         return 0;
 434     ch = (int)value;
 435 #ifndef CHARSET_EBCDIC
 436     if (!isdigit(ch) && ch != ' ')
 437         return 0;
 438 #else
 439     if (ch > os_toascii['9'])
 440         return 0;
 441     if (ch < os_toascii['0'] && ch != os_toascii[' '])
 442         return 0;
 443 #endif
 444     return 1;
 445 }