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 "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_printable(unsigned long value);
  76
  77 /*
  78  * These functions take a string in UTF8, ASCII or multibyte form and a mask
  79  * of permissible ASN1 string types. It then works out the minimal type
  80  * (using the order Printable < IA5 < T61 < BMP < Universal < UTF8) and
  81  * creates a string of the correct type with the supplied data. Yes this is
  82  * horrible: it has to be :-( The 'ncopy' form checks minimum and maximum
  83  * size limits too.
  84  */
  85
  86 int ASN1_mbstring_copy(ASN1_STRING **out, const unsigned char *in, int len,
  87                        int inform, unsigned long mask)
  88 {
  89     return ASN1_mbstring_ncopy(out, in, len, inform, mask, 0, 0);
  90 }
  91
  92 int ASN1_mbstring_ncopy(ASN1_STRING **out, const unsigned char *in, int len,
  93                         int inform, unsigned long mask,
  94                         long minsize, long maxsize)
  95 {
  96     int str_type;
  97     int ret;
  98     char free_out;
  99     int outform, outlen = 0;
 100     ASN1_STRING *dest;
 101     unsigned char *p;
 102     int nchar;
 103     char strbuf[32];
 104     int (*cpyfunc) (unsigned long, void *) = NULL;
 105     if (len == -1)
 106         len = strlen((const char *)in);
 107     if (!mask)
 108         mask = DIRSTRING_TYPE;
 109
 110     /* First do a string check and work out the number of characters */
 111     switch (inform) {
 112
 113     case MBSTRING_BMP:
 114         if (len & 1) {
 115             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY,
 116                     ASN1_R_INVALID_BMPSTRING_LENGTH);
 117             return -1;
 118         }
 119         nchar = len >> 1;
 120         break;
 121
 122     case MBSTRING_UNIV:
 123         if (len & 3) {
 124             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY,
 125                     ASN1_R_INVALID_UNIVERSALSTRING_LENGTH);
 126             return -1;
 127         }
 128         nchar = len >> 2;
 129         break;
 130
 131     case MBSTRING_UTF8:
 132         nchar = 0;
 133         /* This counts the characters and does utf8 syntax checking */
 134         ret = traverse_string(in, len, MBSTRING_UTF8, in_utf8, &nchar);
 135         if (ret < 0) {
 136             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_INVALID_UTF8STRING);
 137             return -1;
 138         }
 139         break;
 140
 141     case MBSTRING_ASC:
 142         nchar = len;
 143         break;
 144
 145     default:
 146         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_UNKNOWN_FORMAT);
 147         return -1;
 148     }
 149
 150     if ((minsize > 0) && (nchar < minsize)) {
 151         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_STRING_TOO_SHORT);
 152         BIO_snprintf(strbuf, sizeof strbuf, "%ld", minsize);
 153         ERR_add_error_data(2, "minsize=", strbuf);
 154         return -1;
 155     }
 156
 157     if ((maxsize > 0) && (nchar > maxsize)) {
 158         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_STRING_TOO_LONG);
 159         BIO_snprintf(strbuf, sizeof strbuf, "%ld", maxsize);
 160         ERR_add_error_data(2, "maxsize=", strbuf);
 161         return -1;
 162     }
 163
 164     /* Now work out minimal type (if any) */
 165     if (traverse_string(in, len, inform, type_str, &mask) < 0) {
 166         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ASN1_R_ILLEGAL_CHARACTERS);
 167         return -1;
 168     }
 169
 170     /* Now work out output format and string type */
 171     outform = MBSTRING_ASC;
 172     if (mask & B_ASN1_PRINTABLESTRING)
 173         str_type = V_ASN1_PRINTABLESTRING;
 174     else if (mask & B_ASN1_IA5STRING)
 175         str_type = V_ASN1_IA5STRING;
 176     else if (mask & B_ASN1_T61STRING)
 177         str_type = V_ASN1_T61STRING;
 178     else if (mask & B_ASN1_BMPSTRING) {
 179         str_type = V_ASN1_BMPSTRING;
 180         outform = MBSTRING_BMP;
 181     } else if (mask & B_ASN1_UNIVERSALSTRING) {
 182         str_type = V_ASN1_UNIVERSALSTRING;
 183         outform = MBSTRING_UNIV;
 184     } else {
 185         str_type = V_ASN1_UTF8STRING;
 186         outform = MBSTRING_UTF8;
 187     }
 188     if (!out)
 189         return str_type;
 190     if (*out) {
 191         free_out = 0;
 192         dest = *out;
 193         OPENSSL_free(dest->data);
 194         dest->data = NULL;
 195         dest->length = 0;
 196         dest->type = str_type;
 197     } else {
 198         free_out = 1;
 199         dest = ASN1_STRING_type_new(str_type);
 200         if (!dest) {
 201             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 202             return -1;
 203         }
 204         *out = dest;
 205     }
 206     /* If both the same type just copy across */
 207     if (inform == outform) {
 208         if (!ASN1_STRING_set(dest, in, len)) {
 209             ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 210             return -1;
 211         }
 212         return str_type;
 213     }
 214
 215     /* Work out how much space the destination will need */
 216     switch (outform) {
 217     case MBSTRING_ASC:
 218         outlen = nchar;
 219         cpyfunc = cpy_asc;
 220         break;
 221
 222     case MBSTRING_BMP:
 223         outlen = nchar << 1;
 224         cpyfunc = cpy_bmp;
 225         break;
 226
 227     case MBSTRING_UNIV:
 228         outlen = nchar << 2;
 229         cpyfunc = cpy_univ;
 230         break;
 231
 232     case MBSTRING_UTF8:
 233         outlen = 0;
 234         traverse_string(in, len, inform, out_utf8, &outlen);
 235         cpyfunc = cpy_utf8;
 236         break;
 237     }
 238     if ((p = OPENSSL_malloc(outlen + 1)) == NULL) {
 239         if (free_out)
 240             ASN1_STRING_free(dest);
 241         ASN1err(ASN1_F_ASN1_MBSTRING_NCOPY, ERR_R_MALLOC_FAILURE);
 242         return -1;
 243     }
 244     dest->length = outlen;
 245     dest->data = p;
 246     p[outlen] = 0;
 247     traverse_string(in, len, inform, cpyfunc, &p);
 248     return str_type;
 249 }
 250
 251 /*
 252  * This function traverses a string and passes the value of each character to
 253  * an optional function along with a void * argument.
 254  */
 255
 256 static int traverse_string(const unsigned char *p, int len, int inform,
 257                            int (*rfunc) (unsigned long value, void *in),
 258                            void *arg)
 259 {
 260     unsigned long value;
 261     int ret;
 262     while (len) {
 263         if (inform == MBSTRING_ASC) {
 264             value = *p++;
 265             len--;
 266         } else if (inform == MBSTRING_BMP) {
 267             value = *p++ << 8;
 268             value |= *p++;
 269             len -= 2;
 270         } else if (inform == MBSTRING_UNIV) {
 271             value = ((unsigned long)*p++) << 24;
 272             value |= ((unsigned long)*p++) << 16;
 273             value |= *p++ << 8;
 274             value |= *p++;
 275             len -= 4;
 276         } else {
 277             ret = UTF8_getc(p, len, &value);
 278             if (ret < 0)
 279                 return -1;
 280             len -= ret;
 281             p += ret;
 282         }
 283         if (rfunc) {
 284             ret = rfunc(value, arg);
 285             if (ret <= 0)
 286                 return ret;
 287         }
 288     }
 289     return 1;
 290 }
 291
 292 /* Various utility functions for traverse_string */
 293
 294 /* Just count number of characters */
 295
 296 static int in_utf8(unsigned long value, void *arg)
 297 {
 298     int *nchar;
 299     nchar = arg;
 300     (*nchar)++;
 301     return 1;
 302 }
 303
 304 /* Determine size of output as a UTF8 String */
 305
 306 static int out_utf8(unsigned long value, void *arg)
 307 {
 308     int *outlen;
 309     outlen = arg;
 310     *outlen += UTF8_putc(NULL, -1, value);
 311     return 1;
 312 }
 313
 314 /*
 315  * Determine the "type" of a string: check each character against a supplied
 316  * "mask".
 317  */
 318
 319 static int type_str(unsigned long value, void *arg)
 320 {
 321     unsigned long types;
 322     types = *((unsigned long *)arg);
 323     if ((types & B_ASN1_PRINTABLESTRING) && !is_printable(value))
 324         types &= ~B_ASN1_PRINTABLESTRING;
 325     if ((types & B_ASN1_IA5STRING) && (value > 127))
 326         types &= ~B_ASN1_IA5STRING;
 327     if ((types & B_ASN1_T61STRING) && (value > 0xff))
 328         types &= ~B_ASN1_T61STRING;
 329     if ((types & B_ASN1_BMPSTRING) && (value > 0xffff))
 330         types &= ~B_ASN1_BMPSTRING;
 331     if (!types)
 332         return -1;
 333     *((unsigned long *)arg) = types;
 334     return 1;
 335 }
 336
 337 /* Copy one byte per character ASCII like strings */
 338
 339 static int cpy_asc(unsigned long value, void *arg)
 340 {
 341     unsigned char **p, *q;
 342     p = arg;
 343     q = *p;
 344     *q = (unsigned char)value;
 345     (*p)++;
 346     return 1;
 347 }
 348
 349 /* Copy two byte per character BMPStrings */
 350
 351 static int cpy_bmp(unsigned long value, void *arg)
 352 {
 353     unsigned char **p, *q;
 354     p = arg;
 355     q = *p;
 356     *q++ = (unsigned char)((value >> 8) & 0xff);
 357     *q = (unsigned char)(value & 0xff);
 358     *p += 2;
 359     return 1;
 360 }
 361
 362 /* Copy four byte per character UniversalStrings */
 363
 364 static int cpy_univ(unsigned long value, void *arg)
 365 {
 366     unsigned char **p, *q;
 367     p = arg;
 368     q = *p;
 369     *q++ = (unsigned char)((value >> 24) & 0xff);
 370     *q++ = (unsigned char)((value >> 16) & 0xff);
 371     *q++ = (unsigned char)((value >> 8) & 0xff);
 372     *q = (unsigned char)(value & 0xff);
 373     *p += 4;
 374     return 1;
 375 }
 376
 377 /* Copy to a UTF8String */
 378
 379 static int cpy_utf8(unsigned long value, void *arg)
 380 {
 381     unsigned char **p;
 382     int ret;
 383     p = arg;
 384     /* We already know there is enough room so pass 0xff as the length */
 385     ret = UTF8_putc(*p, 0xff, value);
 386     *p += ret;
 387     return 1;
 388 }
 389
 390 /* Return 1 if the character is permitted in a PrintableString */
 391 static int is_printable(unsigned long value)
 392 {
 393     int ch;
 394     if (value > 0x7f)
 395         return 0;
 396     ch = (int)value;
 397     /*
 398      * Note: we can't use 'isalnum' because certain accented characters may
 399      * count as alphanumeric in some environments.
 400      */
 401 #ifndef CHARSET_EBCDIC
 402     if ((ch >= 'a') && (ch <= 'z'))
 403         return 1;
 404     if ((ch >= 'A') && (ch <= 'Z'))
 405         return 1;
 406     if ((ch >= '0') && (ch <= '9'))
 407         return 1;
 408     if ((ch == ' ') || strchr("'()+,-./:=?", ch))
 409         return 1;
 410 #else                           /* CHARSET_EBCDIC */
 411     if ((ch >= os_toascii['a']) && (ch <= os_toascii['z']))
 412         return 1;
 413     if ((ch >= os_toascii['A']) && (ch <= os_toascii['Z']))
 414         return 1;
 415     if ((ch >= os_toascii['0']) && (ch <= os_toascii['9']))
 416         return 1;
 417     if ((ch == os_toascii[' ']) || strchr("'()+,-./:=?", os_toebcdic[ch]))
 418         return 1;
 419 #endif                          /* CHARSET_EBCDIC */
 420     return 0;
 421 }