rpcs3/rpcs3/Emu/CPU/CPUTranslator.h
2018-02-08 21:10:03 +03:00

750 lines
16 KiB
C++

#pragma once
#ifdef LLVM_AVAILABLE
#include "restore_new.h"
#ifdef _MSC_VER
#pragma warning(push, 0)
#endif
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Module.h"
#ifdef _MSC_VER
#pragma warning(pop)
#endif
#include "define_new_memleakdetect.h"
#include "../Utilities/types.h"
#include "../Utilities/StrFmt.h"
#include "../Utilities/BEType.h"
#include "../Utilities/BitField.h"
#include <unordered_map>
#include <map>
#include <unordered_set>
#include <set>
#include <array>
#include <vector>
template <typename T = void>
struct llvm_value_t
{
static_assert(std::is_same<T, void>::value, "llvm_value_t<> error: unknown type");
using type = void;
static constexpr uint esize = 0;
static constexpr bool is_int = false;
static constexpr bool is_sint = false;
static constexpr bool is_uint = false;
static constexpr bool is_float = false;
static constexpr uint is_vector = false;
static constexpr uint is_pointer = false;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getVoidTy(context);
}
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
return value;
}
llvm::Value* value;
// llvm_value_t() = default;
// llvm_value_t(llvm::Value* value)
// : value(value)
// {
// }
};
template <>
struct llvm_value_t<bool> : llvm_value_t<void>
{
using type = bool;
using base = llvm_value_t<void>;
using base::base;
static constexpr uint esize = 1;
static constexpr uint is_int = true;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getInt1Ty(context);
}
};
template <>
struct llvm_value_t<char> : llvm_value_t<void>
{
using type = char;
using base = llvm_value_t<void>;
using base::base;
static constexpr uint esize = 8;
static constexpr bool is_int = true;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getInt8Ty(context);
}
};
template <>
struct llvm_value_t<s8> : llvm_value_t<char>
{
using type = s8;
using base = llvm_value_t<char>;
using base::base;
static constexpr bool is_sint = true;
};
template <>
struct llvm_value_t<u8> : llvm_value_t<char>
{
using type = u8;
using base = llvm_value_t<char>;
using base::base;
static constexpr bool is_uint = true;
};
template <>
struct llvm_value_t<s16> : llvm_value_t<s8>
{
using type = s16;
using base = llvm_value_t<s8>;
using base::base;
static constexpr uint esize = 16;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getInt16Ty(context);
}
};
template <>
struct llvm_value_t<u16> : llvm_value_t<s16>
{
using type = u16;
using base = llvm_value_t<s16>;
using base::base;
static constexpr bool is_sint = false;
static constexpr bool is_uint = true;
};
template <>
struct llvm_value_t<s32> : llvm_value_t<s8>
{
using type = s32;
using base = llvm_value_t<s8>;
using base::base;
static constexpr uint esize = 32;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getInt32Ty(context);
}
};
template <>
struct llvm_value_t<u32> : llvm_value_t<s32>
{
using type = u32;
using base = llvm_value_t<s32>;
using base::base;
static constexpr bool is_sint = false;
static constexpr bool is_uint = true;
};
template <>
struct llvm_value_t<s64> : llvm_value_t<s8>
{
using type = s64;
using base = llvm_value_t<s8>;
using base::base;
static constexpr uint esize = 64;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getInt64Ty(context);
}
};
template <>
struct llvm_value_t<u64> : llvm_value_t<s64>
{
using type = u64;
using base = llvm_value_t<s64>;
using base::base;
static constexpr bool is_sint = false;
static constexpr bool is_uint = true;
};
template <>
struct llvm_value_t<s128> : llvm_value_t<s8>
{
using type = s128;
using base = llvm_value_t<s8>;
using base::base;
static constexpr uint esize = 128;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getIntNTy(context, 128);
}
};
template <>
struct llvm_value_t<u128> : llvm_value_t<s128>
{
using type = u128;
using base = llvm_value_t<s128>;
using base::base;
static constexpr bool is_sint = false;
static constexpr bool is_uint = true;
};
template <>
struct llvm_value_t<f32> : llvm_value_t<void>
{
using type = f32;
using base = llvm_value_t<void>;
using base::base;
static constexpr uint esize = 32;
static constexpr bool is_float = true;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getFloatTy(context);
}
};
template <>
struct llvm_value_t<f64> : llvm_value_t<void>
{
using type = f64;
using base = llvm_value_t<void>;
using base::base;
static constexpr uint esize = 64;
static constexpr bool is_float = true;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::Type::getDoubleTy(context);
}
};
template <typename T>
struct llvm_value_t<T*> : llvm_value_t<T>
{
static_assert(!std::is_void<T>::value, "llvm_value_t<> error: invalid pointer to void type");
using type = T*;
using base = llvm_value_t<T>;
using base::base;
static constexpr uint is_pointer = llvm_value_t<T>::is_pointer + 1;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm_value_t<T>::get_type(context)->getPointerTo();
}
};
template <typename T, uint N>
struct llvm_value_t<T[N]> : llvm_value_t<T>
{
static_assert(!llvm_value_t<T>::is_vector, "llvm_value_t<> error: invalid multidimensional vector");
static_assert(!llvm_value_t<T>::is_pointer, "llvm_value_t<>: vector of pointers is not allowed");
using type = T[N];
using base = llvm_value_t<T>;
using base::base;
static constexpr uint is_vector = N;
static constexpr uint is_pointer = 0;
static llvm::Type* get_type(llvm::LLVMContext& context)
{
return llvm::VectorType::get(llvm_value_t<T>::get_type(context), N);
}
};
template <typename T, typename A1, typename A2>
struct llvm_add_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint || llvm_value_t<T>::is_float, "llvm_add_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateAdd(v1, v2);
}
if (llvm_value_t<T>::is_float)
{
return ir->CreateFAdd(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_add_t<typename T1::type, T1, T2> operator +(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T, typename A1, typename A2>
struct llvm_sub_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint || llvm_value_t<T>::is_float, "llvm_sub_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateSub(v1, v2);
}
if (llvm_value_t<T>::is_float)
{
return ir->CreateFSub(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_sub_t<typename T1::type, T1, T2> operator -(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T, typename A1, typename A2>
struct llvm_mul_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint || llvm_value_t<T>::is_float, "llvm_mul_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateMul(v1, v2);
}
if (llvm_value_t<T>::is_float)
{
return ir->CreateFMul(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_mul_t<typename T1::type, T1, T2> operator *(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T, typename A1, typename A2>
struct llvm_div_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint || llvm_value_t<T>::is_float, "llvm_div_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_sint)
{
return ir->CreateSDiv(v1, v2);
}
if (llvm_value_t<T>::is_uint)
{
return ir->CreateUDiv(v1, v2);
}
if (llvm_value_t<T>::is_float)
{
return ir->CreateFDiv(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_div_t<typename T1::type, T1, T2> operator /(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T, typename A1>
struct llvm_neg_t
{
using type = T;
A1 a1;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint || llvm_value_t<T>::is_float, "llvm_neg_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateNeg(v1);
}
if (llvm_value_t<T>::is_float)
{
return ir->CreateFNeg(v1);
}
}
};
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::esize>>
inline llvm_neg_t<typename T1::type, T1> operator -(T1 a1)
{
return {a1};
}
// Constant int helper
struct llvm_int_t
{
u64 value;
u64 eval(llvm::IRBuilder<>*) const
{
return value;
}
};
template <typename T, typename A1, typename A2>
struct llvm_shl_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint, "llvm_shl_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_sint)
{
return ir->CreateShl(v1, v2);
}
if (llvm_value_t<T>::is_uint)
{
return ir->CreateShl(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_shl_t<typename T1::type, T1, T2> operator <<(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_shl_t<typename T1::type, T1, llvm_int_t> operator <<(T1 a1, u64 a2)
{
return {a1, llvm_int_t{a2}};
}
template <typename T, typename A1, typename A2>
struct llvm_shr_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_sint || llvm_value_t<T>::is_uint, "llvm_shr_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_sint)
{
return ir->CreateAShr(v1, v2);
}
if (llvm_value_t<T>::is_uint)
{
return ir->CreateLShr(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_shr_t<typename T1::type, T1, T2> operator >>(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_shr_t<typename T1::type, T1, llvm_int_t> operator >>(T1 a1, u64 a2)
{
return {a1, llvm_int_t{a2}};
}
template <typename T, typename A1, typename A2>
struct llvm_and_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_int, "llvm_and_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateAnd(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_and_t<typename T1::type, T1, T2> operator &(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_and_t<typename T1::type, T1, llvm_int_t> operator &(T1 a1, u64 a2)
{
return {a1, llvm_int_t{a2}};
}
template <typename T, typename A1, typename A2>
struct llvm_or_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_int, "llvm_or_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateOr(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_or_t<typename T1::type, T1, T2> operator |(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_or_t<typename T1::type, T1, llvm_int_t> operator |(T1 a1, u64 a2)
{
return {a1, llvm_int_t{a2}};
}
template <typename T, typename A1, typename A2>
struct llvm_xor_t
{
using type = T;
A1 a1;
A2 a2;
static_assert(llvm_value_t<T>::is_int, "llvm_xor_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
const auto v2 = a2.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateXor(v1, v2);
}
}
};
template <typename T1, typename T2, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<std::is_same<typename T1::type, typename T2::type>::value>>
inline llvm_xor_t<typename T1::type, T1, T2> operator ^(T1 a1, T2 a2)
{
return {a1, a2};
}
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_xor_t<typename T1::type, T1, llvm_int_t> operator ^(T1 a1, u64 a2)
{
return {a1, llvm_int_t{a2}};
}
template <typename T, typename A1>
struct llvm_not_t
{
using type = T;
A1 a1;
static_assert(llvm_value_t<T>::is_int, "llvm_not_t<>: invalid type");
llvm::Value* eval(llvm::IRBuilder<>* ir) const
{
const auto v1 = a1.eval(ir);
if (llvm_value_t<T>::is_int)
{
return ir->CreateNot(v1);
}
}
};
template <typename T1, typename = decltype(std::declval<T1>().eval(0)), typename = std::enable_if_t<llvm_value_t<typename T1::type>::is_int>>
inline llvm_not_t<typename T1::type, T1> operator ~(T1 a1)
{
return {a1};
}
class cpu_translator
{
protected:
cpu_translator(llvm::LLVMContext& context, llvm::Module* module, bool is_be);
// LLVM context
llvm::LLVMContext& m_context;
// Module to which all generated code is output to
llvm::Module* const m_module;
// Endianness, affects vector element numbering (TODO)
const bool m_is_be;
// IR builder
llvm::IRBuilder<>* m_ir;
public:
// Convert a C++ type to an LLVM type (TODO: remove)
template <typename T>
llvm::Type* GetType()
{
return llvm_value_t<T>::get_type(m_context);
}
template <typename T>
llvm::Type* get_type()
{
return llvm_value_t<T>::get_type(m_context);
}
template <typename T>
using value_t = llvm_value_t<T>;
template <typename T>
auto eval(T expr)
{
value_t<typename T::type> result;
result.value = expr.eval(m_ir);
return result;
}
// Get unsigned addition carry into the sign bit (s = a + b)
template <typename T>
static inline auto ucarry(T a, T b, T s)
{
return ((a ^ b) & ~s) | (a & b);
}
// Get signed addition overflow into the sign bit (s = a + b)
template <typename T>
static inline auto scarry(T a, T b, T s)
{
return (b ^ s) & ~(a ^ b);
}
// Get signed subtraction overflow into the sign bit (d = a - b)
template <typename T>
static inline auto sborrow(T a, T b, T d)
{
return (a ^ b) & (a ^ d);
}
// Bitwise select (c ? a : b)
template <typename T>
static inline auto merge(T c, T a, T b)
{
return (a & c) | (b & ~c);
}
// Average: (a + b + 1) >> 1
template <typename T>
static inline auto avg(T a, T b)
{
return (a >> 1) + (b >> 1) + ((a | b) & 1);
}
};
#endif