intel8086/emulator.h

#include <iostream>
#include <cstdint>
#include <fstream>

#define MEMORY_SIZE 1048576

union reg16 {
    struct {
        uint8_t low;
        uint8_t high;
    };
    uint16_t value;
};

struct ModRM {
    uint8_t mod; // addressing mode
    uint8_t reg; // source
    uint8_t rm; // dest
};

reg16 AX, BX, CX, DX;  // general regs
uint16_t SI, DI;          // source index, dest index
uint16_t BP, SP;          // base ptr, stack ptr
uint16_t IP;              // instruction ptr
uint16_t CS, DS, ES, SS;  // code segmets reg, data segment reg, extra segment reg, stack segment reg

bool CF; // carry
bool PF; // parity
bool AF; // auxiliary carry
bool ZF; // zero
bool SF; // sign
bool TF; // trap
bool IF; // interrupt
bool DF; // direction
bool OF; // overflow

uint8_t ram[MEMORY_SIZE];

uint32_t physical_addr(uint16_t seg, uint16_t offset);
bool load_binary(char* filename, size_t addr);
ModRM parse_modrm(uint8_t byte);

uint8_t get_reg_byte(uint8_t code);
void set_reg_byte(uint8_t code, uint8_t value);
uint16_t get_reg_word(uint8_t code);
void set_reg_word(uint8_t code, uint16_t value);

uint16_t get_displacement(uint8_t size);
uint32_t calc_effective_addr(ModRM modrm);

void update_flags_add(uint8_t b1, uint8_t b2, uint8_t res);
void update_flags_add16(uint16_t w1, uint16_t w2, uint16_t res);


uint32_t physical_addr(uint16_t seg, uint16_t offset) {
    return ((uint32_t) seg << 4) + offset;
}

bool load_binary(char* filename, size_t addr) {
    std::ifstream file(filename, std::ios::binary);
    if (!file) {
        std::cerr << "failed to open file " << filename << "\n";
        return false;
    }
    
    file.seekg(0, std::ios::end);
    size_t file_size = file.tellg();
    file.seekg(0, std::ios::beg);

    if (file_size > MEMORY_SIZE - addr) {
        std::cerr << "file is too big";
        return false;
    }

    file.read(reinterpret_cast<char*>(ram + addr), file_size);

    if (!file) {
        std::cerr << "failed to read entire file";
        return false;
    }

    std::cout << "file " << filename << " loaded to ram\n";
    return true;
}

ModRM parse_modrm(uint8_t byte) {
    ModRM modrm;
    modrm.mod = (byte >> 6) & 0x3;
    modrm.reg = (byte >> 3) & 0x7;
    modrm.rm = byte & 0x7;
    return modrm;
}

uint8_t get_reg_byte(uint8_t code) {
    switch (code) {
        case 0: return AX.low;
        case 1: return CX.low;
        case 2: return DX.low;
        case 3: return BX.low;
        case 4: return AX.high;
        case 5: return CX.high;
        case 6: return DX.high;
        case 7: return BX.high;
        default:
            std::cerr << "invalid reg code for 8-bit reg";
            return 0;
    }
}

void set_reg_byte(uint8_t code, uint8_t value) {
    switch (code) {
        case 0: AX.low = value; break;
        case 1: CX.low = value; break;
        case 2: DX.low = value; break;
        case 3: BX.low = value; break;
        case 4: AX.high = value; break;
        case 5: CX.high = value; break;
        case 6: DX.high = value; break;
        case 7: BX.high = value; break;
        default:
            std::cerr << "invalid reg code for 8-bit reg";
    }
}

uint16_t get_reg_word(uint8_t code) {
    switch (code) {
        case 0: return AX.value;
        case 1: return CX.value;
        case 2: return DX.value;
        case 3: return BX.value;
        case 4: return SP;
        case 5: return BP;
        case 6: return SI;
        case 7: return DI;
        default:
            std::cerr << "invalid reg code for 8-bit reg";
            return 0;
    }
}

void set_reg_word(uint8_t code, uint16_t value) {
    switch (code)
    {
        case 0: AX.value = value; break;
        case 1: CX.value = value; break;
        case 2: DX.value = value; break;
        case 3: BX.value = value; break;
        case 4: SP = value; break;
        case 5: BP = value; break;
        case 6: SI = value; break;
        case 7: DI = value; break;
        
        default:
            std::cerr << "invalid reg code for 8-bit reg";
            break;
    }
}

uint16_t get_displacement(uint8_t size) {
    uint16_t disp = 0;
    uint32_t addr = physical_addr(CS, IP);

    if (size == 1) { // 8-bit
        disp = static_cast<int8_t>(ram[addr]);
        IP++;
    } else if (size == 2) {
        disp = ram[addr] | (ram[addr + 1] << 8);
        IP += 2;
    }

    return disp;
}

uint32_t calc_effective_addr(ModRM modrm) {
    uint16_t base = 0;
    uint16_t disp = 0;
    bool has_disp = false;

    switch (modrm.rm) {
        case 0: base = BX.value + SI; break;
        case 1: base = BX.value + DI; break;
        case 2: base = BP + SI; break;
        case 3: base = BP + DI; break;
        case 4: base = SI; break;
        case 5: base = DI; break;
        case 6:
            if (modrm.mod == 0) {
                disp = get_displacement(2);
                base = 0;  // Для прямого адреса
            } else base = BP;
            break;
        case 7: base = BX.value; break;
    }

    if (modrm.mod == 1) {
        disp = get_displacement(1);
        has_disp = true;
    } else if (modrm.mod == 2) {
        disp = get_displacement(2);
        has_disp = true;
    }

    return base + (has_disp ? disp : 0);
}

void update_flags_add(uint8_t b1, uint8_t b2, uint8_t res) {
    CF = (static_cast<uint16_t>(b1) + static_cast<uint16_t>(b2)) > 0xFF;
    ZF = res == 0;
    SF = (res & 0x80) != 0;
    
    uint8_t count = 0;
    for (int i = 0; i < 8; ++i)
        count += (res >> i) & 1;
    PF = (count % 2) == 0;

    AF = ((b1 & 0xF) + (b2 & 0xF)) > 0xF;
    OF = (~(b1 ^ b2) & (b2 ^ res) & 0x80) != 0;
}

void update_flags_add16(uint16_t w1, uint16_t w2, uint16_t res) {
    CF = (static_cast<uint32_t>(w1) + static_cast<uint32_t>(w2)) > 0xFFFF;
    ZF = (res == 0);
    SF = (res & 0x8000) != 0;
    
    uint8_t low_byte = res & 0xFF;
    uint8_t count = 0;
    for (int i = 0; i < 8; i++) 
        count += (low_byte >> i) & 1;
    PF = (count % 2 == 0);

    AF = ((w1 & 0xF) + (w2 & 0xF)) > 0xF;
    OF = ((w1 ^ res) & (w2 ^ res) & 0x8000) != 0;
}