mirror of
https://github.com/trezor/trezor-firmware.git
synced 2024-12-27 00:28:10 +00:00
185 lines
7.1 KiB
C
185 lines
7.1 KiB
C
/**
|
|
* Copyright (c) 2013-2014 Tomas Dzetkulic
|
|
* Copyright (c) 2013-2014 Pavol Rusnak
|
|
* Copyright (c) 2016 Alex Beregszaszi
|
|
*
|
|
* 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.
|
|
*/
|
|
|
|
#ifndef __BIGNUM_H__
|
|
#define __BIGNUM_H__
|
|
|
|
#include <stdbool.h>
|
|
#include <stddef.h>
|
|
#include <stdint.h>
|
|
|
|
#include "options.h"
|
|
|
|
#define BN_LIMBS 9
|
|
#define BN_BITS_PER_LIMB 29
|
|
#define BN_BASE (1u << BN_BITS_PER_LIMB)
|
|
#define BN_LIMB_MASK ((1u << BN_BITS_PER_LIMB) - 1)
|
|
#define BN_EXTRA_BITS (32 - BN_BITS_PER_LIMB)
|
|
#define BN_BITS_LAST_LIMB (256 - (BN_LIMBS - 1) * BN_BITS_PER_LIMB)
|
|
|
|
// Represents the number sum([val[i] * 2**(29*i) for i in range(9))
|
|
typedef struct {
|
|
uint32_t val[BN_LIMBS];
|
|
} bignum256;
|
|
|
|
// Represents the number sum([val[i] * 2**(29*i) for i in range(18))
|
|
typedef struct {
|
|
uint32_t val[2 * BN_LIMBS];
|
|
} bignum512;
|
|
|
|
static inline uint32_t read_be(const uint8_t *data) {
|
|
return (((uint32_t)data[0]) << 24) | (((uint32_t)data[1]) << 16) |
|
|
(((uint32_t)data[2]) << 8) | (((uint32_t)data[3]));
|
|
}
|
|
|
|
static inline void write_be(uint8_t *data, uint32_t x) {
|
|
data[0] = x >> 24;
|
|
data[1] = x >> 16;
|
|
data[2] = x >> 8;
|
|
data[3] = x;
|
|
}
|
|
|
|
static inline uint32_t read_le(const uint8_t *data) {
|
|
return (((uint32_t)data[3]) << 24) | (((uint32_t)data[2]) << 16) |
|
|
(((uint32_t)data[1]) << 8) | (((uint32_t)data[0]));
|
|
}
|
|
|
|
static inline void write_le(uint8_t *data, uint32_t x) {
|
|
data[3] = x >> 24;
|
|
data[2] = x >> 16;
|
|
data[1] = x >> 8;
|
|
data[0] = x;
|
|
}
|
|
|
|
void bn_copy_lower(const bignum512 *x, bignum256 *y);
|
|
void bn_read_be(const uint8_t *in_number, bignum256 *out_number);
|
|
void bn_read_be_512(const uint8_t *in_number, bignum512 *out_number);
|
|
void bn_write_be(const bignum256 *in_number, uint8_t *out_number);
|
|
void bn_read_le(const uint8_t *in_number, bignum256 *out_number);
|
|
void bn_write_le(const bignum256 *in_number, uint8_t *out_number);
|
|
void bn_read_uint32(uint32_t in_number, bignum256 *out_number);
|
|
void bn_read_uint64(uint64_t in_number, bignum256 *out_number);
|
|
int bn_bitcount(const bignum256 *x);
|
|
unsigned int bn_digitcount(const bignum256 *x);
|
|
void bn_zero(bignum256 *x);
|
|
void bn_one(bignum256 *x);
|
|
int bn_is_zero(const bignum256 *x);
|
|
int bn_is_one(const bignum256 *x);
|
|
int bn_is_less(const bignum256 *x, const bignum256 *y);
|
|
int bn_is_equal(const bignum256 *x, const bignum256 *y);
|
|
void bn_cmov(bignum256 *res, volatile uint32_t cond, const bignum256 *truecase,
|
|
const bignum256 *falsecase);
|
|
void bn_cnegate(volatile uint32_t cond, bignum256 *x, const bignum256 *prime);
|
|
void bn_lshift(bignum256 *x);
|
|
void bn_rshift(bignum256 *x);
|
|
void bn_setbit(bignum256 *x, uint16_t i);
|
|
void bn_clearbit(bignum256 *x, uint16_t i);
|
|
uint32_t bn_testbit(const bignum256 *x, uint16_t i);
|
|
void bn_xor(bignum256 *res, const bignum256 *x, const bignum256 *y);
|
|
void bn_mult_half(bignum256 *x, const bignum256 *prime);
|
|
void bn_mult_k(bignum256 *x, uint8_t k, const bignum256 *prime);
|
|
void bn_mod(bignum256 *x, const bignum256 *prime);
|
|
void bn_multiply(const bignum256 *k, bignum256 *x, const bignum256 *prime);
|
|
void bn_reduce(bignum512 *x, const bignum256 *prime);
|
|
void bn_fast_mod(bignum256 *x, const bignum256 *prime);
|
|
void bn_power_mod(const bignum256 *x, const bignum256 *e,
|
|
const bignum256 *prime, bignum256 *res);
|
|
void bn_sqrt(bignum256 *x, const bignum256 *prime);
|
|
uint32_t inverse_mod_power_two(uint32_t a, uint32_t n);
|
|
void bn_divide_base(bignum256 *x, const bignum256 *prime);
|
|
void bn_normalize(bignum256 *x);
|
|
void bn_add(bignum256 *x, const bignum256 *y);
|
|
void bn_addmod(bignum256 *x, const bignum256 *y, const bignum256 *prime);
|
|
void bn_addi(bignum256 *x, uint32_t y);
|
|
void bn_subi(bignum256 *x, uint32_t y, const bignum256 *prime);
|
|
void bn_subtractmod(const bignum256 *x, const bignum256 *y, bignum256 *res,
|
|
const bignum256 *prime);
|
|
void bn_subtract(const bignum256 *x, const bignum256 *y, bignum256 *res);
|
|
int bn_legendre(const bignum256 *x, const bignum256 *prime);
|
|
void bn_long_division(bignum256 *x, uint32_t d, bignum256 *q, uint32_t *r);
|
|
void bn_divmod58(bignum256 *x, uint32_t *r);
|
|
void bn_divmod1000(bignum256 *x, uint32_t *r);
|
|
void bn_inverse(bignum256 *x, const bignum256 *prime);
|
|
size_t bn_format(const bignum256 *amount, const char *prefix,
|
|
const char *suffix, unsigned int decimals, int exponent,
|
|
bool trailing, char thousands, char *output,
|
|
size_t output_length);
|
|
|
|
// Returns (uint32_t) in_number
|
|
// Assumes in_number < 2**32
|
|
// Assumes in_number is normalized
|
|
static inline uint32_t bn_write_uint32(const bignum256 *in_number) {
|
|
return in_number->val[0] | (in_number->val[1] << BN_BITS_PER_LIMB);
|
|
}
|
|
|
|
// Returns (uint64_t) in_number
|
|
// Assumes in_number < 2**64
|
|
// Assumes in_number is normalized
|
|
static inline uint64_t bn_write_uint64(const bignum256 *in_number) {
|
|
uint64_t acc;
|
|
acc = in_number->val[2];
|
|
acc <<= BN_BITS_PER_LIMB;
|
|
acc |= in_number->val[1];
|
|
acc <<= BN_BITS_PER_LIMB;
|
|
acc |= in_number->val[0];
|
|
return acc;
|
|
}
|
|
|
|
// y = x
|
|
static inline void bn_copy(const bignum256 *x, bignum256 *y) { *y = *x; }
|
|
|
|
// Returns x % 2 == 0
|
|
static inline int bn_is_even(const bignum256 *x) {
|
|
return (x->val[0] & 1) == 0;
|
|
}
|
|
|
|
// Returns x % 2 == 0
|
|
static inline int bn_is_odd(const bignum256 *x) { return (x->val[0] & 1) == 1; }
|
|
|
|
static inline size_t bn_format_uint64(uint64_t amount, const char *prefix,
|
|
const char *suffix, unsigned int decimals,
|
|
int exponent, bool trailing,
|
|
char thousands, char *output,
|
|
size_t output_length) {
|
|
bignum256 bn_amount;
|
|
bn_read_uint64(amount, &bn_amount);
|
|
|
|
return bn_format(&bn_amount, prefix, suffix, decimals, exponent, trailing,
|
|
thousands, output, output_length);
|
|
}
|
|
|
|
static inline size_t bn_format_amount(uint64_t amount, const char *prefix,
|
|
const char *suffix, unsigned int decimals,
|
|
char *output, size_t output_length) {
|
|
return bn_format_uint64(amount, prefix, suffix, decimals, 0, false, ',',
|
|
output, output_length);
|
|
}
|
|
|
|
#if USE_BN_PRINT
|
|
void bn_print(const bignum256 *x);
|
|
void bn_print_raw(const bignum256 *x);
|
|
#endif
|
|
|
|
#endif
|