Home | History | Annotate | Download | only in base64
      1 /* Copyright (C) 1995-1998 Eric Young (eay (at) cryptsoft.com)
      2  * All rights reserved.
      3  *
      4  * This package is an SSL implementation written
      5  * by Eric Young (eay (at) cryptsoft.com).
      6  * The implementation was written so as to conform with Netscapes SSL.
      7  *
      8  * This library is free for commercial and non-commercial use as long as
      9  * the following conditions are aheared to.  The following conditions
     10  * apply to all code found in this distribution, be it the RC4, RSA,
     11  * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
     12  * included with this distribution is covered by the same copyright terms
     13  * except that the holder is Tim Hudson (tjh (at) cryptsoft.com).
     14  *
     15  * Copyright remains Eric Young's, and as such any Copyright notices in
     16  * the code are not to be removed.
     17  * If this package is used in a product, Eric Young should be given attribution
     18  * as the author of the parts of the library used.
     19  * This can be in the form of a textual message at program startup or
     20  * in documentation (online or textual) provided with the package.
     21  *
     22  * Redistribution and use in source and binary forms, with or without
     23  * modification, are permitted provided that the following conditions
     24  * are met:
     25  * 1. Redistributions of source code must retain the copyright
     26  *    notice, this list of conditions and the following disclaimer.
     27  * 2. Redistributions in binary form must reproduce the above copyright
     28  *    notice, this list of conditions and the following disclaimer in the
     29  *    documentation and/or other materials provided with the distribution.
     30  * 3. All advertising materials mentioning features or use of this software
     31  *    must display the following acknowledgement:
     32  *    "This product includes cryptographic software written by
     33  *     Eric Young (eay (at) cryptsoft.com)"
     34  *    The word 'cryptographic' can be left out if the rouines from the library
     35  *    being used are not cryptographic related :-).
     36  * 4. If you include any Windows specific code (or a derivative thereof) from
     37  *    the apps directory (application code) you must include an acknowledgement:
     38  *    "This product includes software written by Tim Hudson (tjh (at) cryptsoft.com)"
     39  *
     40  * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
     41  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     42  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     43  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
     44  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     45  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     46  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     47  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     48  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     49  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     50  * SUCH DAMAGE.
     51  *
     52  * The licence and distribution terms for any publically available version or
     53  * derivative of this code cannot be changed.  i.e. this code cannot simply be
     54  * copied and put under another distribution licence
     55  * [including the GNU Public Licence.] */
     56 
     57 #include <openssl/base64.h>
     58 
     59 #include <assert.h>
     60 #include <limits.h>
     61 #include <string.h>
     62 
     63 #include <openssl/type_check.h>
     64 
     65 #include "../internal.h"
     66 
     67 
     68 // constant_time_lt_args_8 behaves like |constant_time_lt_8| but takes |uint8_t|
     69 // arguments for a slightly simpler implementation.
     70 static inline uint8_t constant_time_lt_args_8(uint8_t a, uint8_t b) {
     71   crypto_word_t aw = a;
     72   crypto_word_t bw = b;
     73   // |crypto_word_t| is larger than |uint8_t|, so |aw| and |bw| have the same
     74   // MSB. |aw| < |bw| iff MSB(|aw| - |bw|) is 1.
     75   return constant_time_msb_w(aw - bw);
     76 }
     77 
     78 // constant_time_in_range_8 returns |CONSTTIME_TRUE_8| if |min| <= |a| <= |max|
     79 // and |CONSTTIME_FALSE_8| otherwise.
     80 static inline uint8_t constant_time_in_range_8(uint8_t a, uint8_t min,
     81                                                uint8_t max) {
     82   a -= min;
     83   return constant_time_lt_args_8(a, max - min + 1);
     84 }
     85 
     86 // Encoding.
     87 
     88 static uint8_t conv_bin2ascii(uint8_t a) {
     89   // Since PEM is sometimes used to carry private keys, we encode base64 data
     90   // itself in constant-time.
     91   a &= 0x3f;
     92   uint8_t ret = constant_time_select_8(constant_time_eq_8(a, 62), '+', '/');
     93   ret =
     94       constant_time_select_8(constant_time_lt_args_8(a, 62), a - 52 + '0', ret);
     95   ret =
     96       constant_time_select_8(constant_time_lt_args_8(a, 52), a - 26 + 'a', ret);
     97   ret = constant_time_select_8(constant_time_lt_args_8(a, 26), a + 'A', ret);
     98   return ret;
     99 }
    100 
    101 OPENSSL_COMPILE_ASSERT(sizeof(((EVP_ENCODE_CTX *)(NULL))->data) % 3 == 0,
    102                        data_length_must_be_multiple_of_base64_chunk_size);
    103 
    104 int EVP_EncodedLength(size_t *out_len, size_t len) {
    105   if (len + 2 < len) {
    106     return 0;
    107   }
    108   len += 2;
    109   len /= 3;
    110 
    111   if (((len << 2) >> 2) != len) {
    112     return 0;
    113   }
    114   len <<= 2;
    115 
    116   if (len + 1 < len) {
    117     return 0;
    118   }
    119   len++;
    120 
    121   *out_len = len;
    122   return 1;
    123 }
    124 
    125 void EVP_EncodeInit(EVP_ENCODE_CTX *ctx) {
    126   OPENSSL_memset(ctx, 0, sizeof(EVP_ENCODE_CTX));
    127 }
    128 
    129 void EVP_EncodeUpdate(EVP_ENCODE_CTX *ctx, uint8_t *out, int *out_len,
    130                       const uint8_t *in, size_t in_len) {
    131   size_t total = 0;
    132 
    133   *out_len = 0;
    134   if (in_len == 0) {
    135     return;
    136   }
    137 
    138   assert(ctx->data_used < sizeof(ctx->data));
    139 
    140   if (sizeof(ctx->data) - ctx->data_used > in_len) {
    141     OPENSSL_memcpy(&ctx->data[ctx->data_used], in, in_len);
    142     ctx->data_used += (unsigned)in_len;
    143     return;
    144   }
    145 
    146   if (ctx->data_used != 0) {
    147     const size_t todo = sizeof(ctx->data) - ctx->data_used;
    148     OPENSSL_memcpy(&ctx->data[ctx->data_used], in, todo);
    149     in += todo;
    150     in_len -= todo;
    151 
    152     size_t encoded = EVP_EncodeBlock(out, ctx->data, sizeof(ctx->data));
    153     ctx->data_used = 0;
    154 
    155     out += encoded;
    156     *(out++) = '\n';
    157     *out = '\0';
    158 
    159     total = encoded + 1;
    160   }
    161 
    162   while (in_len >= sizeof(ctx->data)) {
    163     size_t encoded = EVP_EncodeBlock(out, in, sizeof(ctx->data));
    164     in += sizeof(ctx->data);
    165     in_len -= sizeof(ctx->data);
    166 
    167     out += encoded;
    168     *(out++) = '\n';
    169     *out = '\0';
    170 
    171     if (total + encoded + 1 < total) {
    172       *out_len = 0;
    173       return;
    174     }
    175 
    176     total += encoded + 1;
    177   }
    178 
    179   if (in_len != 0) {
    180     OPENSSL_memcpy(ctx->data, in, in_len);
    181   }
    182 
    183   ctx->data_used = (unsigned)in_len;
    184 
    185   if (total > INT_MAX) {
    186     // We cannot signal an error, but we can at least avoid making *out_len
    187     // negative.
    188     total = 0;
    189   }
    190   *out_len = (int)total;
    191 }
    192 
    193 void EVP_EncodeFinal(EVP_ENCODE_CTX *ctx, uint8_t *out, int *out_len) {
    194   if (ctx->data_used == 0) {
    195     *out_len = 0;
    196     return;
    197   }
    198 
    199   size_t encoded = EVP_EncodeBlock(out, ctx->data, ctx->data_used);
    200   out[encoded++] = '\n';
    201   out[encoded] = '\0';
    202   ctx->data_used = 0;
    203 
    204   // ctx->data_used is bounded by sizeof(ctx->data), so this does not
    205   // overflow.
    206   assert(encoded <= INT_MAX);
    207   *out_len = (int)encoded;
    208 }
    209 
    210 size_t EVP_EncodeBlock(uint8_t *dst, const uint8_t *src, size_t src_len) {
    211   uint32_t l;
    212   size_t remaining = src_len, ret = 0;
    213 
    214   while (remaining) {
    215     if (remaining >= 3) {
    216       l = (((uint32_t)src[0]) << 16L) | (((uint32_t)src[1]) << 8L) | src[2];
    217       *(dst++) = conv_bin2ascii(l >> 18L);
    218       *(dst++) = conv_bin2ascii(l >> 12L);
    219       *(dst++) = conv_bin2ascii(l >> 6L);
    220       *(dst++) = conv_bin2ascii(l);
    221       remaining -= 3;
    222     } else {
    223       l = ((uint32_t)src[0]) << 16L;
    224       if (remaining == 2) {
    225         l |= ((uint32_t)src[1] << 8L);
    226       }
    227 
    228       *(dst++) = conv_bin2ascii(l >> 18L);
    229       *(dst++) = conv_bin2ascii(l >> 12L);
    230       *(dst++) = (remaining == 1) ? '=' : conv_bin2ascii(l >> 6L);
    231       *(dst++) = '=';
    232       remaining = 0;
    233     }
    234     ret += 4;
    235     src += 3;
    236   }
    237 
    238   *dst = '\0';
    239   return ret;
    240 }
    241 
    242 
    243 // Decoding.
    244 
    245 int EVP_DecodedLength(size_t *out_len, size_t len) {
    246   if (len % 4 != 0) {
    247     return 0;
    248   }
    249 
    250   *out_len = (len / 4) * 3;
    251   return 1;
    252 }
    253 
    254 void EVP_DecodeInit(EVP_ENCODE_CTX *ctx) {
    255   OPENSSL_memset(ctx, 0, sizeof(EVP_ENCODE_CTX));
    256 }
    257 
    258 static uint8_t base64_ascii_to_bin(uint8_t a) {
    259   // Since PEM is sometimes used to carry private keys, we decode base64 data
    260   // itself in constant-time.
    261   const uint8_t is_upper = constant_time_in_range_8(a, 'A', 'Z');
    262   const uint8_t is_lower = constant_time_in_range_8(a, 'a', 'z');
    263   const uint8_t is_digit = constant_time_in_range_8(a, '0', '9');
    264   const uint8_t is_plus = constant_time_eq_8(a, '+');
    265   const uint8_t is_slash = constant_time_eq_8(a, '/');
    266   const uint8_t is_equals = constant_time_eq_8(a, '=');
    267 
    268   uint8_t ret = 0xff;  // 0xff signals invalid.
    269   ret = constant_time_select_8(is_upper, a - 'A', ret);       // [0,26)
    270   ret = constant_time_select_8(is_lower, a - 'a' + 26, ret);  // [26,52)
    271   ret = constant_time_select_8(is_digit, a - '0' + 52, ret);  // [52,62)
    272   ret = constant_time_select_8(is_plus, 62, ret);
    273   ret = constant_time_select_8(is_slash, 63, ret);
    274   // Padding maps to zero, to be further handled by the caller.
    275   ret = constant_time_select_8(is_equals, 0, ret);
    276   return ret;
    277 }
    278 
    279 // base64_decode_quad decodes a single quad (i.e. four characters) of base64
    280 // data and writes up to three bytes to |out|. It sets |*out_num_bytes| to the
    281 // number of bytes written, which will be less than three if the quad ended
    282 // with padding.  It returns one on success or zero on error.
    283 static int base64_decode_quad(uint8_t *out, size_t *out_num_bytes,
    284                               const uint8_t *in) {
    285   const uint8_t a = base64_ascii_to_bin(in[0]);
    286   const uint8_t b = base64_ascii_to_bin(in[1]);
    287   const uint8_t c = base64_ascii_to_bin(in[2]);
    288   const uint8_t d = base64_ascii_to_bin(in[3]);
    289   if (a == 0xff || b == 0xff || c == 0xff || d == 0xff) {
    290     return 0;
    291   }
    292 
    293   const uint32_t v = ((uint32_t)a) << 18 | ((uint32_t)b) << 12 |
    294                      ((uint32_t)c) << 6 | (uint32_t)d;
    295 
    296   const unsigned padding_pattern = (in[0] == '=') << 3 |
    297                                    (in[1] == '=') << 2 |
    298                                    (in[2] == '=') << 1 |
    299                                    (in[3] == '=');
    300 
    301   switch (padding_pattern) {
    302     case 0:
    303       // The common case of no padding.
    304       *out_num_bytes = 3;
    305       out[0] = v >> 16;
    306       out[1] = v >> 8;
    307       out[2] = v;
    308       break;
    309 
    310     case 1:  // xxx=
    311       *out_num_bytes = 2;
    312       out[0] = v >> 16;
    313       out[1] = v >> 8;
    314       break;
    315 
    316     case 3:  // xx==
    317       *out_num_bytes = 1;
    318       out[0] = v >> 16;
    319       break;
    320 
    321     default:
    322       return 0;
    323   }
    324 
    325   return 1;
    326 }
    327 
    328 int EVP_DecodeUpdate(EVP_ENCODE_CTX *ctx, uint8_t *out, int *out_len,
    329                      const uint8_t *in, size_t in_len) {
    330   *out_len = 0;
    331 
    332   if (ctx->error_encountered) {
    333     return -1;
    334   }
    335 
    336   size_t bytes_out = 0, i;
    337   for (i = 0; i < in_len; i++) {
    338     const char c = in[i];
    339     switch (c) {
    340       case ' ':
    341       case '\t':
    342       case '\r':
    343       case '\n':
    344         continue;
    345     }
    346 
    347     if (ctx->eof_seen) {
    348       ctx->error_encountered = 1;
    349       return -1;
    350     }
    351 
    352     ctx->data[ctx->data_used++] = c;
    353     if (ctx->data_used == 4) {
    354       size_t num_bytes_resulting;
    355       if (!base64_decode_quad(out, &num_bytes_resulting, ctx->data)) {
    356         ctx->error_encountered = 1;
    357         return -1;
    358       }
    359 
    360       ctx->data_used = 0;
    361       bytes_out += num_bytes_resulting;
    362       out += num_bytes_resulting;
    363 
    364       if (num_bytes_resulting < 3) {
    365         ctx->eof_seen = 1;
    366       }
    367     }
    368   }
    369 
    370   if (bytes_out > INT_MAX) {
    371     ctx->error_encountered = 1;
    372     *out_len = 0;
    373     return -1;
    374   }
    375   *out_len = (int)bytes_out;
    376 
    377   if (ctx->eof_seen) {
    378     return 0;
    379   }
    380 
    381   return 1;
    382 }
    383 
    384 int EVP_DecodeFinal(EVP_ENCODE_CTX *ctx, uint8_t *out, int *out_len) {
    385   *out_len = 0;
    386   if (ctx->error_encountered || ctx->data_used != 0) {
    387     return -1;
    388   }
    389 
    390   return 1;
    391 }
    392 
    393 int EVP_DecodeBase64(uint8_t *out, size_t *out_len, size_t max_out,
    394                      const uint8_t *in, size_t in_len) {
    395   *out_len = 0;
    396 
    397   if (in_len % 4 != 0) {
    398     return 0;
    399   }
    400 
    401   size_t max_len;
    402   if (!EVP_DecodedLength(&max_len, in_len) ||
    403       max_out < max_len) {
    404     return 0;
    405   }
    406 
    407   size_t i, bytes_out = 0;
    408   for (i = 0; i < in_len; i += 4) {
    409     size_t num_bytes_resulting;
    410 
    411     if (!base64_decode_quad(out, &num_bytes_resulting, &in[i])) {
    412       return 0;
    413     }
    414 
    415     bytes_out += num_bytes_resulting;
    416     out += num_bytes_resulting;
    417     if (num_bytes_resulting != 3 && i != in_len - 4) {
    418       return 0;
    419     }
    420   }
    421 
    422   *out_len = bytes_out;
    423   return 1;
    424 }
    425 
    426 int EVP_DecodeBlock(uint8_t *dst, const uint8_t *src, size_t src_len) {
    427   // Trim spaces and tabs from the beginning of the input.
    428   while (src_len > 0) {
    429     if (src[0] != ' ' && src[0] != '\t') {
    430       break;
    431     }
    432 
    433     src++;
    434     src_len--;
    435   }
    436 
    437   // Trim newlines, spaces and tabs from the end of the line.
    438   while (src_len > 0) {
    439     switch (src[src_len-1]) {
    440       case ' ':
    441       case '\t':
    442       case '\r':
    443       case '\n':
    444         src_len--;
    445         continue;
    446     }
    447 
    448     break;
    449   }
    450 
    451   size_t dst_len;
    452   if (!EVP_DecodedLength(&dst_len, src_len) ||
    453       dst_len > INT_MAX ||
    454       !EVP_DecodeBase64(dst, &dst_len, dst_len, src, src_len)) {
    455     return -1;
    456   }
    457 
    458   // EVP_DecodeBlock does not take padding into account, so put the
    459   // NULs back in... so the caller can strip them back out.
    460   while (dst_len % 3 != 0) {
    461     dst[dst_len++] = '\0';
    462   }
    463   assert(dst_len <= INT_MAX);
    464 
    465   return (int)dst_len;
    466 }
    467