1
0
mirror of https://github.com/trezor/trezor-firmware.git synced 2024-11-05 06:59:01 +00:00
trezor-firmware/crypto/base58.c

222 lines
6.7 KiB
C
Raw Normal View History

/**
2014-12-23 00:17:58 +00:00
* Copyright (c) 2012-2014 Luke Dashjr
* Copyright (c) 2013-2014 Pavol Rusnak
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included
* in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES
* OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
* OTHER DEALINGS IN THE SOFTWARE.
*/
#include "base58.h"
2014-12-23 00:17:58 +00:00
#include <stdbool.h>
#include <string.h>
#include "memzero.h"
#include "ripemd160.h"
#include "sha2.h"
const char b58digits_ordered[] =
"123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz";
2018-07-05 00:10:34 +00:00
const int8_t b58digits_map[] = {
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, 1, 2, 3, 4, 5, 6, 7,
8, -1, -1, -1, -1, -1, -1, -1, 9, 10, 11, 12, 13, 14, 15, 16, -1, 17, 18,
19, 20, 21, -1, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, -1, -1, -1, -1,
-1, -1, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, -1, 44, 45, 46, 47, 48,
49, 50, 51, 52, 53, 54, 55, 56, 57, -1, -1, -1, -1, -1,
2014-12-23 00:17:58 +00:00
};
2019-09-29 08:46:21 +00:00
typedef uint64_t b58_maxint_t;
typedef uint32_t b58_almostmaxint_t;
#define b58_almostmaxint_bits (sizeof(b58_almostmaxint_t) * 8)
static const b58_almostmaxint_t b58_almostmaxint_mask =
((((b58_maxint_t)1) << b58_almostmaxint_bits) - 1);
// Decodes a null-terminated Base58 string `b58` to binary and writes the result
// at the end of the buffer `bin` of size `*binszp`. On success `*binszp` is set
// to the number of valid bytes at the end of the buffer.
bool b58tobin(void *bin, size_t *binszp, const char *b58) {
size_t binsz = *binszp;
if (binsz == 0) {
return false;
}
const unsigned char *b58u = (const unsigned char *)b58;
unsigned char *binu = bin;
2019-09-29 08:46:21 +00:00
size_t outisz =
(binsz + sizeof(b58_almostmaxint_t) - 1) / sizeof(b58_almostmaxint_t);
b58_almostmaxint_t outi[outisz];
b58_maxint_t t = 0;
b58_almostmaxint_t c = 0;
size_t i = 0, j = 0;
2019-09-29 08:46:21 +00:00
uint8_t bytesleft = binsz % sizeof(b58_almostmaxint_t);
b58_almostmaxint_t zeromask =
bytesleft ? (b58_almostmaxint_mask << (bytesleft * 8)) : 0;
unsigned zerocount = 0;
2014-12-23 00:17:58 +00:00
2019-09-29 08:46:21 +00:00
size_t b58sz = strlen(b58);
2014-12-23 00:17:58 +00:00
memzero(outi, sizeof(outi));
2014-12-23 00:17:58 +00:00
// Leading zeros, just count
for (i = 0; i < b58sz && b58u[i] == '1'; ++i) ++zerocount;
2014-12-23 00:17:58 +00:00
for (; i < b58sz; ++i) {
if (b58u[i] & 0x80)
// High-bit set on invalid digit
return false;
if (b58digits_map[b58u[i]] == -1)
// Invalid base58 digit
return false;
c = (unsigned)b58digits_map[b58u[i]];
for (j = outisz; j--;) {
2019-09-29 08:46:21 +00:00
t = ((b58_maxint_t)outi[j]) * 58 + c;
c = t >> b58_almostmaxint_bits;
outi[j] = t & b58_almostmaxint_mask;
}
if (c)
// Output number too big (carry to the next int32)
return false;
if (outi[0] & zeromask)
// Output number too big (last int32 filled too far)
return false;
}
2014-12-23 00:17:58 +00:00
j = 0;
2019-09-29 08:46:21 +00:00
if (bytesleft) {
for (i = bytesleft; i > 0; --i) {
*(binu++) = (outi[0] >> (8 * (i - 1))) & 0xff;
}
++j;
}
2014-12-23 00:17:58 +00:00
for (; j < outisz; ++j) {
2019-09-29 08:46:21 +00:00
for (i = sizeof(*outi); i > 0; --i) {
*(binu++) = (outi[j] >> (8 * (i - 1))) & 0xff;
}
}
2014-12-23 00:17:58 +00:00
// locate the most significant byte
binu = bin;
for (i = 0; i < binsz; ++i) {
if (binu[i]) break;
}
2019-09-29 08:46:21 +00:00
// prepend the correct number of null-bytes
if (zerocount > i) {
/* result too large */
return false;
}
*binszp = binsz - i + zerocount;
2014-12-23 00:17:58 +00:00
return true;
2014-12-23 00:17:58 +00:00
}
int b58check(const void *bin, size_t binsz, HasherType hasher_type,
const char *base58str) {
unsigned char buf[32] = {0};
const uint8_t *binc = bin;
unsigned i = 0;
if (binsz < 4) return -4;
hasher_Raw(hasher_type, bin, binsz - 4, buf);
if (memcmp(&binc[binsz - 4], buf, 4)) return -1;
2014-12-23 00:17:58 +00:00
// Check number of zeros is correct AFTER verifying checksum (to avoid
// possibility of accessing base58str beyond the end)
for (i = 0; binc[i] == '\0' && base58str[i] == '1'; ++i) {
} // Just finding the end of zeros, nothing to do in loop
if (binc[i] == '\0' || base58str[i] == '1') return -3;
2014-12-23 00:17:58 +00:00
return binc[0];
2014-12-23 00:17:58 +00:00
}
bool b58enc(char *b58, size_t *b58sz, const void *data, size_t binsz) {
const uint8_t *bin = data;
int carry = 0;
size_t i = 0, j = 0, high = 0, zcount = 0;
size_t size = 0;
2014-12-23 00:17:58 +00:00
2019-09-29 08:46:21 +00:00
while (zcount < binsz && !bin[zcount]) ++zcount;
2014-12-23 00:17:58 +00:00
size = (binsz - zcount) * 138 / 100 + 1;
uint8_t buf[size];
memzero(buf, size);
2014-12-23 00:17:58 +00:00
2019-09-29 08:46:21 +00:00
for (i = zcount, high = size - 1; i < binsz; ++i, high = j) {
for (carry = bin[i], j = size - 1; (j > high) || carry; --j) {
carry += 256 * buf[j];
buf[j] = carry % 58;
carry /= 58;
2019-09-29 08:46:21 +00:00
if (!j) {
// Otherwise j wraps to maxint which is > high
break;
}
}
}
2014-12-23 00:17:58 +00:00
2019-09-29 08:46:21 +00:00
for (j = 0; j < size && !buf[j]; ++j)
;
2014-12-23 00:17:58 +00:00
if (*b58sz <= zcount + size - j) {
*b58sz = zcount + size - j + 1;
return false;
}
2014-12-23 00:17:58 +00:00
if (zcount) memset(b58, '1', zcount);
2019-09-29 08:46:21 +00:00
for (i = zcount; j < size; ++i, ++j) b58[i] = b58digits_ordered[buf[j]];
b58[i] = '\0';
*b58sz = i + 1;
2014-12-23 00:17:58 +00:00
return true;
2014-12-23 00:17:58 +00:00
}
int base58_encode_check(const uint8_t *data, int datalen,
HasherType hasher_type, char *str, int strsize) {
if (datalen > 128) {
return 0;
}
uint8_t buf[datalen + 32];
memset(buf, 0, sizeof(buf));
uint8_t *hash = buf + datalen;
memcpy(buf, data, datalen);
hasher_Raw(hasher_type, data, datalen, hash);
size_t res = strsize;
bool success = b58enc(str, &res, buf, datalen + 4);
memzero(buf, sizeof(buf));
return success ? res : 0;
}
int base58_decode_check(const char *str, HasherType hasher_type, uint8_t *data,
int datalen) {
if (datalen > 128) {
return 0;
}
uint8_t d[datalen + 4];
memset(d, 0, sizeof(d));
size_t res = datalen + 4;
if (b58tobin(d, &res, str) != true) {
return 0;
}
uint8_t *nd = d + datalen + 4 - res;
if (b58check(nd, res, hasher_type, str) < 0) {
return 0;
}
memcpy(data, nd, res - 4);
return res - 4;
}