crypto/asn1/a_mbstr.c

   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
  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
  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             ERR_raise(ERR_LIB_ASN1, ERR_R_ASN1_LIB);
 157             return -1;
 158         }
 159         return str_type;
 160     }
 161
 162     /* Work out how much space the destination will need */
 163     switch (outform) {
 164     case MBSTRING_ASC:
 165         outlen = nchar;
 166         cpyfunc = cpy_asc;
 167         break;
 168
 169     case MBSTRING_BMP:
 170         outlen = nchar << 1;
 171         cpyfunc = cpy_bmp;
 172         break;
 173
 174     case MBSTRING_UNIV:
 175         outlen = nchar << 2;
 176         cpyfunc = cpy_univ;
 177         break;
 178
 179     case MBSTRING_UTF8:
 180         outlen = 0;
 181         traverse_string(in, len, inform, out_utf8, &outlen);
 182         cpyfunc = cpy_utf8;
 183         break;
 184     }
 185     if ((p = OPENSSL_malloc(outlen + 1)) == NULL) {
 186         if (free_out)
 187             ASN1_STRING_free(dest);
 188         return -1;
 189     }
 190     dest->length = outlen;
 191     dest->data = p;
 192     p[outlen] = 0;
 193     traverse_string(in, len, inform, cpyfunc, &p);
 194     return str_type;
 195 }
 196
 197 /*
 198  * This function traverses a string and passes the value of each character to
 199  * an optional function along with a void * argument.
 200  */
 201
 202 static int traverse_string(const unsigned char *p, int len, int inform,
 203                            int (*rfunc) (unsigned long value, void *in),
 204                            void *arg)
 205 {
 206     unsigned long value;
 207     int ret;
 208     while (len) {
 209         if (inform == MBSTRING_ASC) {
 210             value = *p++;
 211             len--;
 212         } else if (inform == MBSTRING_BMP) {
 213             value = *p++ << 8;
 214             value |= *p++;
 215             len -= 2;
 216         } else if (inform == MBSTRING_UNIV) {
 217             value = ((unsigned long)*p++) << 24;
 218             value |= ((unsigned long)*p++) << 16;
 219             value |= *p++ << 8;
 220             value |= *p++;
 221             len -= 4;
 222         } else {
 223             ret = UTF8_getc(p, len, &value);
 224             if (ret < 0)
 225                 return -1;
 226             len -= ret;
 227             p += ret;
 228         }
 229         if (rfunc) {
 230             ret = rfunc(value, arg);
 231             if (ret <= 0)
 232                 return ret;
 233         }
 234     }
 235     return 1;
 236 }
 237
 238 /* Various utility functions for traverse_string */
 239
 240 /* Just count number of characters */
 241
 242 static int in_utf8(unsigned long value, void *arg)
 243 {
 244     int *nchar;
 245
 246     if (!is_unicode_valid(value))
 247         return -2;
 248     nchar = arg;
 249     (*nchar)++;
 250     return 1;
 251 }
 252
 253 /* Determine size of output as a UTF8 String */
 254
 255 static int out_utf8(unsigned long value, void *arg)
 256 {
 257     int *outlen, len;
 258
 259     len = UTF8_putc(NULL, -1, value);
 260     if (len <= 0)
 261         return len;
 262     outlen = arg;
 263     *outlen += len;
 264     return 1;
 265 }
 266
 267 /*
 268  * Determine the "type" of a string: check each character against a supplied
 269  * "mask".
 270  */
 271
 272 static int type_str(unsigned long value, void *arg)
 273 {
 274     unsigned long types = *((unsigned long *)arg);
 275     const int native = value > INT_MAX ? INT_MAX : ossl_fromascii(value);
 276
 277     if ((types & B_ASN1_NUMERICSTRING) && !(ossl_isdigit(native)
 278                                             || native == ' '))
 279         types &= ~B_ASN1_NUMERICSTRING;
 280     if ((types & B_ASN1_PRINTABLESTRING) && !ossl_isasn1print(native))
 281         types &= ~B_ASN1_PRINTABLESTRING;
 282     if ((types & B_ASN1_IA5STRING) && !ossl_isascii(native))
 283         types &= ~B_ASN1_IA5STRING;
 284     if ((types & B_ASN1_T61STRING) && (value > 0xff))
 285         types &= ~B_ASN1_T61STRING;
 286     if ((types & B_ASN1_BMPSTRING) && (value > 0xffff))
 287         types &= ~B_ASN1_BMPSTRING;
 288     if ((types & B_ASN1_UTF8STRING) && !is_unicode_valid(value))
 289         types &= ~B_ASN1_UTF8STRING;
 290     if (!types)
 291         return -1;
 292     *((unsigned long *)arg) = types;
 293     return 1;
 294 }
 295
 296 /* Copy one byte per character ASCII like strings */
 297
 298 static int cpy_asc(unsigned long value, void *arg)
 299 {
 300     unsigned char **p, *q;
 301     p = arg;
 302     q = *p;
 303     *q = (unsigned char)value;
 304     (*p)++;
 305     return 1;
 306 }
 307
 308 /* Copy two byte per character BMPStrings */
 309
 310 static int cpy_bmp(unsigned long value, void *arg)
 311 {
 312     unsigned char **p, *q;
 313     p = arg;
 314     q = *p;
 315     *q++ = (unsigned char)((value >> 8) & 0xff);
 316     *q = (unsigned char)(value & 0xff);
 317     *p += 2;
 318     return 1;
 319 }
 320
 321 /* Copy four byte per character UniversalStrings */
 322
 323 static int cpy_univ(unsigned long value, void *arg)
 324 {
 325     unsigned char **p, *q;
 326     p = arg;
 327     q = *p;
 328     *q++ = (unsigned char)((value >> 24) & 0xff);
 329     *q++ = (unsigned char)((value >> 16) & 0xff);
 330     *q++ = (unsigned char)((value >> 8) & 0xff);
 331     *q = (unsigned char)(value & 0xff);
 332     *p += 4;
 333     return 1;
 334 }
 335
 336 /* Copy to a UTF8String */
 337
 338 static int cpy_utf8(unsigned long value, void *arg)
 339 {
 340     unsigned char **p;
 341     int ret;
 342     p = arg;
 343     /* We already know there is enough room so pass 0xff as the length */
 344     ret = UTF8_putc(*p, 0xff, value);
 345     *p += ret;
 346     return 1;
 347 }