ladybird/AK/UFixedBigInt.h

656 lines
23 KiB
C++

/*
* Copyright (c) 2021, Leon Albrecht <leon2002.la@gmail.com>
* Copyright (c) 2023, Dan Klishch <danilklishch@gmail.com>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#pragma once
#include <AK/BigIntBase.h>
#include <AK/BuiltinWrappers.h>
#include <AK/Checked.h>
#include <AK/Concepts.h>
#include <AK/Endian.h>
#include <AK/Format.h>
#include <AK/NumericLimits.h>
#include <AK/StdLibExtraDetails.h>
#include <AK/StdLibExtras.h>
#include <AK/StringBuilder.h>
namespace AK {
namespace Detail {
// As noted near the declaration of StaticStorage, bit_size is more like a hint for a storage size.
// The effective bit size is `sizeof(StaticStorage<...>) * 8`. It is a programmer's responsibility
// to ensure that the hinted bit_size is always greater than the actual integer size.
// That said, do not use unaligned (bit_size % 64 != 0) `UFixedBigInt`s if you do not know what you
// are doing.
template<size_t bit_size, typename Storage = StaticStorage<false, bit_size>>
class UFixedBigInt;
// ===== Concepts =====
template<typename T>
constexpr inline size_t assumed_bit_size = 0;
template<>
constexpr inline size_t assumed_bit_size<IntegerWrapper> = bit_width<int>;
template<size_t bit_size>
constexpr inline size_t assumed_bit_size<UFixedBigInt<bit_size>> = bit_size;
template<BuiltInUFixedInt T>
constexpr inline size_t assumed_bit_size<T> = bit_width<T>;
template<typename T>
concept ConvertibleToUFixedInt = (assumed_bit_size<T> != 0);
template<typename T>
concept UFixedInt = (ConvertibleToUFixedInt<T> && !IsSame<T, IntegerWrapper>);
template<typename T>
concept NotBuiltInUFixedInt = (UFixedInt<T> && !BuiltInUFixedInt<T>);
// ===== UFixedBigInt itself =====
template<size_t bit_size>
constexpr auto& get_storage_of(UFixedBigInt<bit_size>& value) { return value.m_data; }
template<size_t bit_size>
constexpr auto& get_storage_of(UFixedBigInt<bit_size> const& value) { return value.m_data; }
template<typename Operand1, typename Operand2, typename Result>
constexpr void mul_internal(Operand1 const& operand1, Operand2 const& operand2, Result& result)
{
StorageOperations::baseline_mul(operand1, operand2, result, g_null_allocator);
}
template<size_t dividend_size, size_t divisor_size, bool restore_remainder>
constexpr void div_mod_internal( // Include AK/UFixedBigIntDivision.h to use UFixedBigInt division
StaticStorage<false, dividend_size> const& dividend,
StaticStorage<false, divisor_size> const& divisor,
StaticStorage<false, dividend_size>& quotient,
StaticStorage<false, divisor_size>& remainder);
template<size_t bit_size, typename Storage>
class UFixedBigInt {
constexpr static size_t static_size = Storage::static_size;
constexpr static size_t part_size = static_size / 2;
using UFixedBigIntPart = Conditional<part_size * word_size <= 64, u64, UFixedBigInt<part_size * word_size>>;
public:
constexpr UFixedBigInt() = default;
explicit constexpr UFixedBigInt(IntegerWrapper value) { StorageOperations::copy(value.m_data, m_data); }
consteval UFixedBigInt(int value)
{
StorageOperations::copy(IntegerWrapper(value).m_data, m_data);
}
template<UFixedInt T>
requires(sizeof(T) > sizeof(Storage)) explicit constexpr UFixedBigInt(T const& value)
{
StorageOperations::copy(get_storage_of(value), m_data);
}
template<UFixedInt T>
requires(sizeof(T) <= sizeof(Storage)) constexpr UFixedBigInt(T const& value)
{
StorageOperations::copy(get_storage_of(value), m_data);
}
constexpr UFixedBigInt(UFixedBigIntPart const& low, UFixedBigIntPart const& high)
requires(static_size % 2 == 0)
{
decltype(auto) low_storage = get_storage_of(low);
decltype(auto) high_storage = get_storage_of(high);
for (size_t i = 0; i < part_size; ++i)
m_data[i] = low_storage[i];
for (size_t i = 0; i < part_size; ++i)
m_data[i + part_size] = high_storage[i];
}
template<UFixedInt T, size_t n>
requires((assumed_bit_size<T> * n) <= bit_size) constexpr UFixedBigInt(const T (&value)[n])
{
size_t offset = 0;
for (size_t i = 0; i < n; ++i) {
if (offset % word_size == 0) {
// Aligned initialization (i. e. u256 from two u128)
decltype(auto) storage = get_storage_of(value[i]);
for (size_t i = 0; i < storage.size(); ++i)
m_data[i + offset / word_size] = storage[i];
} else if (offset % word_size == 32 && IsSame<T, u32>) {
// u32 vector initialization on 64-bit platforms
m_data[offset / word_size] |= static_cast<DoubleWord>(value[i]) << 32;
} else {
VERIFY_NOT_REACHED();
}
offset += assumed_bit_size<T>;
}
for (size_t i = (offset + word_size - 1) / word_size; i < m_data.size(); ++i)
m_data[i] = 0;
}
// Casts & parts extraction
template<NotBuiltInUFixedInt T>
constexpr explicit operator T() const
{
T result;
StorageOperations::copy(m_data, result.m_data);
return result;
}
template<BuiltInUFixedInt T>
requires(sizeof(T) <= sizeof(NativeWord)) constexpr explicit operator T() const
{
return m_data[0];
}
template<BuiltInUFixedInt T>
requires(sizeof(T) == sizeof(DoubleWord)) constexpr explicit operator T() const
{
return (static_cast<DoubleWord>(m_data[1]) << word_size) + m_data[0];
}
constexpr UFixedBigIntPart low() const
requires(static_size % 2 == 0)
{
if constexpr (part_size == 1) {
return m_data[0];
} else if constexpr (IsSame<UFixedBigIntPart, DoubleWord>) {
return m_data[0] + (static_cast<DoubleWord>(m_data[1]) << word_size);
} else {
UFixedBigInt<part_size * word_size> result;
StorageOperations::copy(m_data, result.m_data);
return result;
}
}
constexpr UFixedBigIntPart high() const
requires(static_size % 2 == 0)
{
if constexpr (part_size == 1) {
return m_data[part_size];
} else if constexpr (IsSame<UFixedBigIntPart, DoubleWord>) {
return m_data[part_size] + (static_cast<DoubleWord>(m_data[part_size + 1]) << word_size);
} else {
UFixedBigInt<part_size * word_size> result;
StorageOperations::copy(m_data, result.m_data, part_size);
return result;
}
}
Bytes bytes()
{
return Bytes(reinterpret_cast<u8*>(this), sizeof(Storage));
}
ReadonlyBytes bytes() const
{
return ReadonlyBytes(reinterpret_cast<u8 const*>(this), sizeof(Storage));
}
constexpr UnsignedStorageSpan span()
{
return { m_data.data(), static_size };
}
constexpr UnsignedStorageReadonlySpan span() const
{
return { m_data.data(), static_size };
}
// Binary utils
constexpr size_t popcnt() const
{
size_t result = 0;
for (size_t i = 0; i < m_data.size(); ++i)
result += popcount(m_data[i]);
return result;
}
constexpr size_t ctz() const
{
size_t result = 0;
for (size_t i = 0; i < m_data.size(); ++i) {
if (m_data[i]) {
result += count_trailing_zeroes(m_data[i]);
break;
} else {
result += word_size;
}
}
return result;
}
constexpr size_t clz() const
{
size_t result = 0;
for (size_t i = m_data.size(); i--;) {
if (m_data[i]) {
result += count_leading_zeroes(m_data[i]);
break;
} else {
result += word_size;
}
}
return result + bit_size - word_size * static_size;
}
// Comparisons
constexpr bool operator!() const
{
bool result = true;
for (size_t i = 0; i < m_data.size(); ++i)
result &= !m_data[i];
return result;
}
constexpr explicit operator bool() const
{
bool result = false;
for (size_t i = 0; i < m_data.size(); ++i)
result |= m_data[i];
return result;
}
constexpr bool operator==(UFixedInt auto const& other) const
{
return StorageOperations::compare(m_data, get_storage_of(other), true) == 0;
}
constexpr bool operator==(IntegerWrapper other) const
{
return StorageOperations::compare(m_data, get_storage_of(other), true) == 0;
}
constexpr int operator<=>(UFixedInt auto const& other) const
{
return StorageOperations::compare(m_data, get_storage_of(other), false);
}
constexpr int operator<=>(IntegerWrapper other) const
{
return StorageOperations::compare(m_data, get_storage_of(other), false);
}
#define DEFINE_STANDARD_BINARY_OPERATOR(op, function) \
constexpr auto operator op(UFixedInt auto const& other) const \
{ \
auto func = [](auto&& a, auto&& b, auto&& c) { function(a, b, c); }; \
return do_standard_binary_operation(other, func); \
} \
\
constexpr auto operator op(IntegerWrapper other) const \
{ \
auto func = [](auto&& a, auto&& b, auto&& c) { function(a, b, c); }; \
return do_standard_binary_operation(other, func); \
}
#define DEFINE_STANDARD_COMPOUND_ASSIGNMENT(op, function) \
constexpr auto& operator op(UFixedInt auto const& other) \
{ \
auto func = [](auto&& a, auto&& b, auto&& c) { function(a, b, c); }; \
do_standard_compound_assignment(other, func); \
return *this; \
} \
\
constexpr auto& operator op(IntegerWrapper other) \
{ \
auto func = [](auto&& a, auto&& b, auto&& c) { function(a, b, c); }; \
do_standard_compound_assignment(other, func); \
return *this; \
}
// Binary operators
DEFINE_STANDARD_BINARY_OPERATOR(^, StorageOperations::compute_bitwise<StorageOperations::Bitwise::XOR>)
DEFINE_STANDARD_BINARY_OPERATOR(&, StorageOperations::compute_bitwise<StorageOperations::Bitwise::AND>)
DEFINE_STANDARD_BINARY_OPERATOR(|, StorageOperations::compute_bitwise<StorageOperations::Bitwise::OR>)
DEFINE_STANDARD_COMPOUND_ASSIGNMENT(^=, StorageOperations::compute_inplace_bitwise<StorageOperations::Bitwise::XOR>)
DEFINE_STANDARD_COMPOUND_ASSIGNMENT(&=, StorageOperations::compute_inplace_bitwise<StorageOperations::Bitwise::AND>)
DEFINE_STANDARD_COMPOUND_ASSIGNMENT(|=, StorageOperations::compute_inplace_bitwise<StorageOperations::Bitwise::OR>)
constexpr auto operator~() const
{
UFixedBigInt<bit_size> result;
StorageOperations::compute_bitwise<StorageOperations::Bitwise::INVERT>(m_data, m_data, result.m_data);
return result;
}
constexpr auto operator<<(size_t shift) const
{
UFixedBigInt<bit_size> result;
StorageOperations::shift_left(m_data, shift, result.m_data);
return result;
}
constexpr auto& operator<<=(size_t shift)
{
StorageOperations::shift_left(m_data, shift, m_data);
return *this;
}
constexpr auto operator>>(size_t shift) const
{
UFixedBigInt<bit_size> result;
StorageOperations::shift_right(m_data, shift, result.m_data);
return result;
}
constexpr auto& operator>>=(size_t shift)
{
StorageOperations::shift_right(m_data, shift, m_data);
return *this;
}
// Arithmetic
template<UFixedInt T>
constexpr auto addc(T const& other, bool& carry) const
{
UFixedBigInt<max(bit_size, assumed_bit_size<T>)> result;
carry = StorageOperations::add<false>(m_data, get_storage_of(other), result.m_data, carry);
return result;
}
template<UFixedInt T>
constexpr auto subc(T const& other, bool& borrow) const
{
UFixedBigInt<max(bit_size, assumed_bit_size<T>)> result;
borrow = StorageOperations::add<true>(m_data, get_storage_of(other), result.m_data, borrow);
return result;
}
DEFINE_STANDARD_BINARY_OPERATOR(+, StorageOperations::add<false>)
DEFINE_STANDARD_BINARY_OPERATOR(-, StorageOperations::add<true>)
DEFINE_STANDARD_COMPOUND_ASSIGNMENT(+=, StorageOperations::add<false>)
DEFINE_STANDARD_COMPOUND_ASSIGNMENT(-=, StorageOperations::add<true>)
constexpr auto& operator++()
{
StorageOperations::increment<false>(m_data);
return *this;
}
constexpr auto& operator--()
{
StorageOperations::increment<true>(m_data);
return *this;
}
constexpr auto operator++(int)
{
UFixedBigInt<bit_size> result = *this;
StorageOperations::increment<false>(m_data);
return result;
}
constexpr auto operator--(int)
{
UFixedBigInt<bit_size> result = *this;
StorageOperations::increment<true>(m_data);
return result;
}
DEFINE_STANDARD_BINARY_OPERATOR(*, mul_internal)
constexpr auto& operator*=(UFixedInt auto const& other) { return *this = *this * other; }
constexpr auto& operator*=(IntegerWrapper const& other) { return *this = *this * other; }
template<UFixedInt T>
constexpr auto wide_multiply(T const& other) const
{
UFixedBigInt<bit_size + assumed_bit_size<T>> result;
mul_internal(m_data, get_storage_of(other), result.m_data);
return result;
}
template<NotBuiltInUFixedInt T>
constexpr UFixedBigInt<bit_size> div_mod(T const& divisor, T& remainder) const
{
UFixedBigInt<bit_size> quotient;
UFixedBigInt<assumed_bit_size<T>> resulting_remainder;
div_mod_internal<bit_size, assumed_bit_size<T>, true>(m_data, get_storage_of(divisor), get_storage_of(quotient), get_storage_of(resulting_remainder));
remainder = resulting_remainder;
return quotient;
}
template<UFixedInt T>
constexpr auto operator/(T const& other) const
{
UFixedBigInt<bit_size> quotient;
StaticStorage<false, assumed_bit_size<T>> remainder; // unused
div_mod_internal<bit_size, assumed_bit_size<T>, false>(m_data, get_storage_of(other), get_storage_of(quotient), remainder);
return quotient;
}
template<UFixedInt T>
constexpr auto operator%(T const& other) const
{
StaticStorage<false, bit_size> quotient; // unused
UFixedBigInt<assumed_bit_size<T>> remainder;
div_mod_internal<bit_size, assumed_bit_size<T>, true>(m_data, get_storage_of(other), quotient, get_storage_of(remainder));
return remainder;
}
constexpr auto operator/(IntegerWrapper const& other) const { return *this / static_cast<UFixedBigInt<32>>(other); }
constexpr auto operator%(IntegerWrapper const& other) const { return *this % static_cast<UFixedBigInt<32>>(other); }
template<UFixedInt T>
constexpr auto& operator/=(T const& other) { return *this = *this / other; }
constexpr auto& operator/=(IntegerWrapper const& other) { return *this = *this / other; }
template<Unsigned U>
constexpr auto& operator%=(U const& other) { return *this = *this % other; }
constexpr auto& operator%=(IntegerWrapper const& other) { return *this = *this % other; }
// Note: If there ever be need for non side-channel proof sqrt/pow/pow_mod of UFixedBigInt, you
// can restore them from Git history.
#undef DEFINE_STANDARD_BINARY_OPERATOR
#undef DEFINE_STANDARD_COMPOUND_ASSIGNMENT
// These functions are intended to be used in LibCrypto for equality checks without branching.
constexpr bool is_zero_constant_time() const
{
NativeWord fold = 0;
for (size_t i = 0; i < m_data.size(); ++i)
taint_for_optimizer(fold |= m_data[i]);
return !fold;
}
constexpr bool is_equal_to_constant_time(UFixedBigInt<bit_size> other) const
{
NativeWord fold = 0;
for (size_t i = 0; i < m_data.size(); ++i)
taint_for_optimizer(fold |= m_data[i] ^ other.m_data[i]);
return !fold;
}
private:
template<ConvertibleToUFixedInt T, typename Function>
constexpr auto do_standard_binary_operation(T const& other, Function function) const
{
UFixedBigInt<max(bit_size, assumed_bit_size<T>)> result;
function(m_data, get_storage_of(other), result.m_data);
return result;
}
template<ConvertibleToUFixedInt T, typename Function>
constexpr void do_standard_compound_assignment(T const& other, Function function)
{
static_assert(bit_size >= assumed_bit_size<T>, "Requested operation requires integer size to be expanded.");
function(m_data, get_storage_of(other), m_data);
}
template<size_t other_bit_size, typename OtherStorage>
friend class UFixedBigInt;
friend constexpr auto& get_storage_of<bit_size>(UFixedBigInt<bit_size>&);
friend constexpr auto& get_storage_of<bit_size>(UFixedBigInt<bit_size> const&);
Storage m_data;
};
// FIXME: There is a bug in LLVM (https://github.com/llvm/llvm-project/issues/59783) which doesn't
// allow to use the following comparisons.
bool operator==(BuiltInUFixedInt auto const& a, NotBuiltInUFixedInt auto const& b) { return b.operator==(a); }
int operator<=>(BuiltInUFixedInt auto const& a, NotBuiltInUFixedInt auto const& b) { return -b.operator<=>(a); }
bool operator==(IntegerWrapper const& a, NotBuiltInUFixedInt auto const& b) { return b.operator==(a); }
int operator<=>(IntegerWrapper const& a, NotBuiltInUFixedInt auto const& b) { return -b.operator<=>(a); }
}
using Detail::UFixedBigInt;
template<size_t bit_size>
constexpr inline bool IsUnsigned<UFixedBigInt<bit_size>> = true;
template<size_t bit_size>
constexpr inline bool IsSigned<UFixedBigInt<bit_size>> = false;
template<size_t bit_size>
struct NumericLimits<UFixedBigInt<bit_size>> {
using T = UFixedBigInt<bit_size>;
static constexpr T min() { return T {}; }
static constexpr T max() { return --T {}; }
static constexpr bool is_signed() { return false; }
};
template<size_t N>
class LittleEndian<UFixedBigInt<N>> {
template<size_t M>
constexpr static auto byte_swap_if_not_little_endian(UFixedBigInt<M> value)
{
if constexpr (HostIsLittleEndian) {
return value;
} else {
auto words = value.span();
auto front_it = words.begin();
auto ending_half_words = words.slice(ceil_div(words.size(), static_cast<size_t>(2)));
for (size_t i = 0; i < ending_half_words.size(); ++i, ++front_it)
*front_it = convert_between_host_and_little_endian(exchange(ending_half_words[ending_half_words.size() - i - 1], convert_between_host_and_little_endian(*front_it)));
if (words.size() % 2)
words[words.size() / 2] = convert_between_host_and_little_endian(*front_it);
return value;
}
}
public:
constexpr LittleEndian() = default;
constexpr LittleEndian(UFixedBigInt<N> value)
: m_value(byte_swap_if_not_little_endian(value))
{
}
constexpr operator UFixedBigInt<N>() const { return byte_swap_if_not_little_endian(m_value); }
private:
UFixedBigInt<N> m_value { 0 };
};
template<size_t N>
class BigEndian<UFixedBigInt<N>> {
template<size_t M>
constexpr static auto byte_swap_if_not_big_endian(UFixedBigInt<M> value)
{
if constexpr (!HostIsLittleEndian) {
return value;
} else {
auto words = value.span();
auto front_it = words.begin();
auto ending_half_words = words.slice(ceil_div(words.size(), static_cast<size_t>(2)));
for (size_t i = 0; i < ending_half_words.size(); ++i, ++front_it)
*front_it = convert_between_host_and_big_endian(exchange(ending_half_words[ending_half_words.size() - i - 1], convert_between_host_and_big_endian(*front_it)));
if (words.size() % 2)
words[words.size() / 2] = convert_between_host_and_big_endian(*front_it);
return value;
}
}
public:
constexpr BigEndian() = default;
constexpr BigEndian(UFixedBigInt<N> value)
: m_value(byte_swap_if_not_big_endian(value))
{
}
constexpr operator UFixedBigInt<N>() const { return byte_swap_if_not_big_endian(m_value); }
private:
UFixedBigInt<N> m_value { 0 };
};
template<size_t M>
struct Traits<UFixedBigInt<M>> : public GenericTraits<UFixedBigInt<M>> {
static constexpr bool is_trivially_serializable() { return true; }
static constexpr bool is_trivial() { return true; }
};
// ===== Formatting =====
// FIXME: This does not work for size != 2 ** x
template<Detail::NotBuiltInUFixedInt T>
struct Formatter<T> : StandardFormatter {
Formatter() = default;
explicit Formatter(StandardFormatter formatter)
: StandardFormatter(formatter)
{
}
ErrorOr<void> format(FormatBuilder& builder, T const& value)
{
using U = decltype(value.low());
if (m_precision.has_value())
VERIFY_NOT_REACHED();
if (m_mode == Mode::Pointer) {
// these are way to big for a pointer
VERIFY_NOT_REACHED();
}
if (m_mode == Mode::Default)
m_mode = Mode::Hexadecimal;
if (!value.high()) {
Formatter<U> formatter { *this };
return formatter.format(builder, value.low());
}
u8 base = 0;
if (m_mode == Mode::Binary) {
base = 2;
} else if (m_mode == Mode::BinaryUppercase) {
base = 2;
} else if (m_mode == Mode::Octal) {
TODO();
} else if (m_mode == Mode::Decimal) {
TODO();
} else if (m_mode == Mode::Hexadecimal) {
base = 16;
} else if (m_mode == Mode::HexadecimalUppercase) {
base = 16;
} else {
VERIFY_NOT_REACHED();
}
ssize_t width = m_width.value_or(0);
ssize_t lower_length = ceil_div(Detail::assumed_bit_size<U>, (ssize_t)base);
Formatter<U> formatter { *this };
formatter.m_width = max(width - lower_length, (ssize_t)0);
TRY(formatter.format(builder, value.high()));
TRY(builder.put_literal("'"sv));
formatter.m_zero_pad = true;
formatter.m_alternative_form = false;
formatter.m_width = lower_length;
TRY(formatter.format(builder, value.low()));
return {};
}
};
}
// these sizes should suffice for most usecases
using u128 = AK::UFixedBigInt<128>;
using u256 = AK::UFixedBigInt<256>;
using u512 = AK::UFixedBigInt<512>;
using u1024 = AK::UFixedBigInt<1024>;
using u2048 = AK::UFixedBigInt<2048>;
using u4096 = AK::UFixedBigInt<4096>;