1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2025-06-08 05:27:14 +09:00
ladybird/Libraries/LibCrypto/BigInt/UnsignedBigInteger.cpp
devgianlu 4b3715ccba LibCrypto: Replace {Unsigned,Signed}BigInteger impl with LibTomMath
Replace the implementation of maths in `UnsignedBigInteger`
and `SignedBigInteger` with LibTomMath. This gives benefits in terms of
less code to maintain, correctness and speed.

These changes also remove now-unsued methods and improve the error
propagation for functions allocating lots of memory. Additionally, the
new implementation is always trimmed and won't have dangling zeros when
exporting it.
2025-05-23 11:57:21 +02:00

510 lines
15 KiB
C++

/*
* Copyright (c) 2020, Itamar S. <itamar8910@gmail.com>
* Copyright (c) 2022, David Tuin <davidot@serenityos.org>
* Copyright (c) 2025, Altomani Gianluca <altomanigianluca@gmail.com>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <math.h>
#include <tommath.h>
#include <AK/BuiltinWrappers.h>
#include <AK/FloatingPoint.h>
#include <AK/NonnullOwnPtr.h>
#include <AK/StringBuilder.h>
#include <LibCrypto/BigInt/Tommath.h>
#include <LibCrypto/BigInt/UnsignedBigInteger.h>
namespace Crypto {
UnsignedBigInteger::UnsignedBigInteger(u8 const* ptr, size_t length)
{
MP_MUST(mp_init(&m_mp));
MP_MUST(mp_from_ubin(&m_mp, ptr, length));
}
UnsignedBigInteger::UnsignedBigInteger(Vector<u32> const& words)
{
MP_MUST(mp_init(&m_mp));
MP_MUST(mp_unpack(&m_mp, words.size(), MP_LSB_FIRST, sizeof(u32), MP_NATIVE_ENDIAN, 0, words.data()));
}
UnsignedBigInteger::UnsignedBigInteger(double value)
{
// Because this is currently only used for LibJS we VERIFY some preconditions
// also these values don't have a clear BigInteger representation.
VERIFY(!isnan(value));
VERIFY(!isinf(value));
VERIFY(trunc(value) == value);
VERIFY(value >= 0.0);
MP_MUST(mp_init(&m_mp));
MP_MUST(mp_set_double(&m_mp, value));
}
UnsignedBigInteger::UnsignedBigInteger(u64 value)
{
MP_MUST(mp_init(&m_mp));
mp_set_u64(&m_mp, value);
}
UnsignedBigInteger::UnsignedBigInteger(UnsignedBigInteger const& other)
{
MP_MUST(mp_init_copy(&m_mp, &other.m_mp));
}
UnsignedBigInteger& UnsignedBigInteger::operator=(UnsignedBigInteger const& other)
{
if (this == &other)
return *this;
mp_clear(&m_mp);
MP_MUST(mp_init_copy(&m_mp, &other.m_mp));
return *this;
}
UnsignedBigInteger::UnsignedBigInteger()
{
MP_MUST(mp_init(&m_mp));
}
UnsignedBigInteger::~UnsignedBigInteger()
{
mp_clear(&m_mp);
}
size_t UnsignedBigInteger::export_data(Bytes data) const
{
size_t written = 0;
MP_MUST(mp_to_ubin(&m_mp, data.data(), data.size(), &written));
return written;
}
ErrorOr<UnsignedBigInteger> UnsignedBigInteger::from_base(u16 N, StringView str)
{
VERIFY(N <= 36);
if (str.is_empty())
return UnsignedBigInteger(0);
auto buffer = TRY(ByteBuffer::create_zeroed(str.length() + 1));
size_t idx = 0;
for (auto& c : str) {
if (c == '_') {
// Skip underscores
continue;
}
buffer[idx++] = c;
}
UnsignedBigInteger result;
if (mp_read_radix(&result.m_mp, reinterpret_cast<char const*>(buffer.data()), N) != MP_OKAY)
return Error::from_string_literal("Invalid number");
return result;
}
ErrorOr<String> UnsignedBigInteger::to_base(u16 N) const
{
VERIFY(N <= 36);
if (is_zero())
return "0"_string;
int size = 0;
MP_MUST(mp_radix_size(&m_mp, N, &size));
auto buffer = TRY(ByteBuffer::create_zeroed(size));
size_t written = 0;
MP_MUST(mp_to_radix(&m_mp, reinterpret_cast<char*>(buffer.data()), size, &written, N));
return StringView(buffer.bytes().slice(0, written - 1)).to_ascii_lowercase_string();
}
u64 UnsignedBigInteger::to_u64() const
{
return mp_get_u64(&m_mp);
}
double UnsignedBigInteger::to_double(RoundingMode rounding_mode) const
{
// Check if we need to truncate
auto bitlen = mp_count_bits(&m_mp);
if (bitlen <= 53)
return mp_get_double(&m_mp);
if (rounding_mode == RoundingMode::RoundTowardZero) {
UnsignedBigInteger shifted;
// Truncate the lower bits
auto shift = bitlen - 53;
MP_MUST(mp_div_2d(&m_mp, shift, &shifted.m_mp, nullptr));
// Convert to double
return ldexp(mp_get_double(&shifted.m_mp), shift);
}
if (rounding_mode == RoundingMode::IEEERoundAndTiesToEvenMantissa) {
UnsignedBigInteger shifted, remainder, half, lsb;
// Get top 53 bits (truncated)
auto shift = bitlen - 53;
MP_MUST(mp_div_2d(&m_mp, shift, &shifted.m_mp, &remainder.m_mp));
// Check if remainder == 2^(shift - 1)
MP_MUST(mp_2expt(&half.m_mp, shift - 1));
int cmp = mp_cmp(&remainder.m_mp, &half.m_mp);
if (cmp < 0) {
// Round down (truncate)
} else if (cmp > 0) {
// Round up
MP_MUST(mp_add_d(&shifted.m_mp, 1, &shifted.m_mp));
} else {
// Exactly halfway, check even
MP_MUST(mp_mod_2d(&shifted.m_mp, 1, &lsb.m_mp));
if (!mp_iszero(&lsb.m_mp)) {
// It's odd, round to even
MP_MUST(mp_add_d(&shifted.m_mp, 1, &shifted.m_mp));
}
}
// Convert to double
return ldexp(mp_get_double(&shifted.m_mp), shift);
}
VERIFY_NOT_REACHED();
}
Vector<u32> UnsignedBigInteger::words() const
{
auto count = mp_pack_count(&m_mp, 0, sizeof(u32));
Vector<u32> result;
result.resize(count);
size_t written = 0;
MP_MUST(mp_pack(result.data(), count, &written, MP_LSB_FIRST, sizeof(u32), MP_NATIVE_ENDIAN, 0, &m_mp));
result.resize(written);
return result;
}
void UnsignedBigInteger::set_to_0()
{
mp_zero(&m_mp);
m_hash = {};
}
void UnsignedBigInteger::set_to(u64 other)
{
mp_set_u64(&m_mp, other);
m_hash = {};
}
void UnsignedBigInteger::set_to(UnsignedBigInteger const& other)
{
MP_MUST(mp_copy(&other.m_mp, &m_mp));
m_hash = {};
}
bool UnsignedBigInteger::is_zero() const
{
return mp_iszero(&m_mp);
}
bool UnsignedBigInteger::is_odd() const
{
return mp_isodd(&m_mp);
}
size_t UnsignedBigInteger::byte_length() const
{
return mp_ubin_size(&m_mp);
}
size_t UnsignedBigInteger::one_based_index_of_highest_set_bit() const
{
return mp_count_bits(&m_mp);
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::plus(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_add(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
FLATTEN ErrorOr<UnsignedBigInteger> UnsignedBigInteger::minus(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_sub(&m_mp, &other.m_mp, &result.m_mp));
if (mp_isneg(&result.m_mp))
return Error::from_string_literal("Substraction produced a negative result");
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::bitwise_or(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_or(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::bitwise_and(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_and(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::bitwise_xor(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_xor(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
FLATTEN ErrorOr<UnsignedBigInteger> UnsignedBigInteger::bitwise_not_fill_to_one_based_index(size_t index) const
{
if (index == 0)
return UnsignedBigInteger(0);
if (index > NumericLimits<int>::max())
return Error::from_errno(ENOMEM);
UnsignedBigInteger result, mask, temp;
MP_TRY(mp_2expt(&mask.m_mp, index));
MP_TRY(mp_sub_d(&mask.m_mp, 1, &mask.m_mp));
MP_TRY(mp_and(&mask.m_mp, &m_mp, &temp.m_mp));
MP_TRY(mp_xor(&temp.m_mp, &mask.m_mp, &result.m_mp));
return result;
}
FLATTEN ErrorOr<UnsignedBigInteger> UnsignedBigInteger::shift_left(size_t num_bits) const
{
UnsignedBigInteger result;
MP_TRY(mp_mul_2d(&m_mp, num_bits, &result.m_mp));
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::shift_right(size_t num_bits) const
{
UnsignedBigInteger result;
MP_MUST(mp_div_2d(&m_mp, num_bits, &result.m_mp, nullptr));
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::multiplied_by(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_mul(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
FLATTEN UnsignedDivisionResult UnsignedBigInteger::divided_by(UnsignedBigInteger const& divisor) const
{
UnsignedBigInteger quotient;
UnsignedBigInteger remainder;
MP_MUST(mp_div(&m_mp, &divisor.m_mp, &quotient.m_mp, &remainder.m_mp));
return UnsignedDivisionResult { quotient, remainder };
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::pow(u32 exponent) const
{
UnsignedBigInteger result;
MP_MUST(mp_expt_n(&m_mp, exponent, &result.m_mp));
return result;
}
FLATTEN UnsignedBigInteger UnsignedBigInteger::gcd(UnsignedBigInteger const& other) const
{
UnsignedBigInteger result;
MP_MUST(mp_gcd(&m_mp, &other.m_mp, &result.m_mp));
return result;
}
u32 UnsignedBigInteger::hash() const
{
if (m_hash.has_value())
return *m_hash;
auto buffer = MUST(ByteBuffer::create_zeroed(byte_length()));
auto length = export_data(buffer);
m_hash = string_hash(reinterpret_cast<char const*>(buffer.data()), length);
return *m_hash;
}
bool UnsignedBigInteger::operator==(UnsignedBigInteger const& other) const
{
return mp_cmp(&m_mp, &other.m_mp) == MP_EQ;
}
bool UnsignedBigInteger::operator!=(UnsignedBigInteger const& other) const
{
return !(*this == other);
}
bool UnsignedBigInteger::operator<(UnsignedBigInteger const& other) const
{
return mp_cmp(&m_mp, &other.m_mp) == MP_LT;
}
bool UnsignedBigInteger::operator<=(UnsignedBigInteger const& other) const
{
return *this < other || *this == other;
}
bool UnsignedBigInteger::operator>(UnsignedBigInteger const& other) const
{
return *this != other && !(*this < other);
}
bool UnsignedBigInteger::operator>=(UnsignedBigInteger const& other) const
{
return *this > other || *this == other;
}
UnsignedBigInteger::CompareResult UnsignedBigInteger::compare_to_double(double value) const
{
VERIFY(!isnan(value));
if (isinf(value)) {
bool is_positive_infinity = __builtin_isinf_sign(value) > 0;
return is_positive_infinity ? CompareResult::DoubleGreaterThanBigInt : CompareResult::DoubleLessThanBigInt;
}
bool value_is_negative = value < 0;
if (value_is_negative)
return CompareResult::DoubleLessThanBigInt;
// Value is zero.
if (value == 0.0) {
VERIFY(!value_is_negative);
// Either we are also zero or value is certainly less than us.
return is_zero() ? CompareResult::DoubleEqualsBigInt : CompareResult::DoubleLessThanBigInt;
}
// If value is not zero but we are, value must be greater.
if (is_zero())
return CompareResult::DoubleGreaterThanBigInt;
FloatExtractor<double> extractor;
extractor.d = value;
// Value cannot be negative at this point.
VERIFY(extractor.sign == 0);
// Exponent cannot be all set, as then we must be NaN or infinity.
VERIFY(extractor.exponent != (1 << extractor.exponent_bits) - 1);
i32 real_exponent = extractor.exponent - extractor.exponent_bias;
if (real_exponent < 0) {
// value is less than 1, and we cannot be zero so value must be less.
return CompareResult::DoubleLessThanBigInt;
}
u64 bigint_bits_needed = one_based_index_of_highest_set_bit();
VERIFY(bigint_bits_needed > 0);
// Double value is `-1^sign (1.mantissa) * 2^(exponent - bias)` so we need
// `exponent - bias + 1` bit to represent doubles value,
// for example `exponent - bias` = 3, sign = 0 and mantissa = 0 we get
// `-1^0 * 2^3 * 1 = 8` which needs 4 bits to store 8 (0b1000).
u32 double_bits_needed = real_exponent + 1;
// If we need more bits to represent us, we must be of greater value.
if (bigint_bits_needed > double_bits_needed)
return CompareResult::DoubleLessThanBigInt;
// If we need less bits to represent us, we must be of less value.
if (bigint_bits_needed < double_bits_needed)
return CompareResult::DoubleGreaterThanBigInt;
u64 mantissa_bits = extractor.mantissa;
// We add the bit which represents the 1. of the double value calculation.
constexpr u64 mantissa_extended_bit = 1ull << extractor.mantissa_bits;
mantissa_bits |= mantissa_extended_bit;
constexpr u32 bits_in_word = sizeof(u32) * 8;
// Now we shift value to the left virtually, with `exponent - bias` steps
// we then pretend both it and the big int are extended with virtual zeros.
auto next_bigint_word = (bits_in_word - 1 + bigint_bits_needed) / bits_in_word;
auto words = this->words();
VERIFY(next_bigint_word == words.size());
auto msb_in_top_word_index = (bigint_bits_needed - 1) % bits_in_word;
VERIFY(msb_in_top_word_index == (bits_in_word - count_leading_zeroes(words[next_bigint_word - 1]) - 1));
// We will keep the bits which are still valid in the mantissa at the top of mantissa bits.
mantissa_bits <<= 64 - (extractor.mantissa_bits + 1);
auto bits_left_in_mantissa = static_cast<size_t>(extractor.mantissa_bits) + 1;
auto get_next_value_bits = [&](size_t num_bits) -> u32 {
VERIFY(num_bits < 63);
VERIFY(bits_left_in_mantissa > 0);
if (num_bits > bits_left_in_mantissa)
num_bits = bits_left_in_mantissa;
bits_left_in_mantissa -= num_bits;
u64 extracted_bits = mantissa_bits & (((1ull << num_bits) - 1) << (64 - num_bits));
// Now shift the bits down to put the most significant bit on the num_bits position
// this means the rest will be "virtual" zeros.
extracted_bits >>= bits_in_word;
// Now shift away the used bits and fit the result into a Word.
mantissa_bits <<= num_bits;
VERIFY(extracted_bits <= NumericLimits<u32>::max());
return static_cast<u32>(extracted_bits);
};
auto bits_in_next_bigint_word = msb_in_top_word_index + 1;
while (next_bigint_word > 0 && bits_left_in_mantissa > 0) {
u32 bigint_word = words[next_bigint_word - 1];
u32 double_word = get_next_value_bits(bits_in_next_bigint_word);
// For the first bit we have to align it with the top bit of bigint
// and for all the other cases bits_in_next_bigint_word is 32 so this does nothing.
double_word >>= bits_in_word - bits_in_next_bigint_word;
if (bigint_word < double_word)
return CompareResult::DoubleGreaterThanBigInt;
if (bigint_word > double_word)
return CompareResult::DoubleLessThanBigInt;
--next_bigint_word;
bits_in_next_bigint_word = bits_in_word;
}
// If there are still bits left in bigint than any non zero bit means it has greater value.
if (next_bigint_word > 0) {
VERIFY(bits_left_in_mantissa == 0);
while (next_bigint_word > 0) {
if (words[next_bigint_word - 1] != 0)
return CompareResult::DoubleLessThanBigInt;
--next_bigint_word;
}
} else if (bits_left_in_mantissa > 0) {
VERIFY(next_bigint_word == 0);
// Similarly if there are still any bits set in the mantissa it has greater value.
if (mantissa_bits != 0)
return CompareResult::DoubleGreaterThanBigInt;
}
// Otherwise if both don't have bits left or the rest of the bits are zero they are equal.
return CompareResult::DoubleEqualsBigInt;
}
}
ErrorOr<void> AK::Formatter<Crypto::UnsignedBigInteger>::format(FormatBuilder& fmtbuilder, Crypto::UnsignedBigInteger const& value)
{
return Formatter<StringView>::format(fmtbuilder, TRY(value.to_base(10)));
}