Start to implement mov

Makefile

@@ -5,7 +5,8 @@ LDFLAGS =
 SRC = src/main.c \
 	  src/dasm.c \
 	  src/instructions.c \
-	  src/interpretor.c
+	  src/interpretor.c \
+	  src/instructions/memory.c
 
 OBJ = $(SRC:%.c=%.o)
 

src/dasm.c

@@ -34,7 +34,7 @@ int print_rm_operand(u_int8_t *binary, int imm_idx, bool is16bit)
 	if (mod == 0b01) {
 		int8_t disp_v = binary[imm_idx];
 		imm_offset++;
-		snprintf(disp, 20, "%c%x", disp_v < 0 ? '-' : '+', disp_v < 0 ? -disp_v : disp_v);
+		snprintf(disp, sizeof(disp), "%c%x", disp_v < 0 ? '-' : '+', disp_v < 0 ? -disp_v : disp_v);
 	}
 
 	switch (rm)
@@ -62,10 +62,10 @@ unsigned read_size(u_int8_t *binary, unsigned size)
 	return ret;
 }
 
-bool has_reg(instruction_t inst)
+bool has_reg(const instruction_t *inst)
 {
-	for (int i = 0; inst.mode[i] != END; i++) {
+	for (int i = 0; inst->mode[i] != END; i++) {
-		if (inst.mode[i] == REG8 || inst.mode[i] == REG16)
+		if (inst->mode[i] == REG8 || inst->mode[i] == REG16)
 			return true;
 	}
 	return false;
@@ -77,7 +77,7 @@ void print_instruction(unsigned addr, instruction_t inst, unsigned inst_size, u_
 	// if the instruction has already a param in it (ex `in al`), we need to directly add a comma.
 	// `jmp short` is an exception as it is the only instruction with a space in it that is not a parameter.
 	bool need_comma = !last_param && strchr(inst.name, ' ') && inst.opcode != 0xEB;
-	int imm_idx = 1 + (inst.extended != -1 || has_reg(inst));
+	int imm_idx = 1 + (inst.extended != -1 || has_reg(&inst));
 
 	printf("%04x:%s%0*x%-*s", addr, space ? " " : "", inst_size * 2, read_size(binary, inst_size), 14 - inst_size * 2, "");
 	if (last_param)

src/dasm.h

@@ -4,6 +4,8 @@
 #include <sys/types.h>
 #include <stdbool.h>
 
+#define FALLTHROUGHT __attribute__((fallthrough));
+
 int dasm(u_int8_t *binary, unsigned long size);
 int interpret(u_int8_t *binary, unsigned long size);
 
@@ -24,6 +26,10 @@ typedef enum addressing_mode {
 	REG8,
 	// 16bit register
 	REG16,
+	// 8bit register but stored inside the opcode
+	OPREG8,
+	// 16bit register but stored inside the opcode
+	OPREG16,
 	// Register or 8bit of memory
 	R_M8,
 	// Register or 16bit of memory
@@ -45,7 +51,7 @@ typedef struct instruction {
 	const char *name;
 	addressing_mode_t mode[5];
 	int extended;
-	void (*exec)(state_t *state);
+	void (*exec)(const struct instruction *self, state_t *state);
 } instruction_t;
 
 extern const instruction_t instructions[];
@@ -55,4 +61,4 @@ extern const instruction_t invalid_instruction;
 instruction_t parse_inst(u_int8_t *binary, unsigned long size);
 unsigned get_inst_size(instruction_t inst, u_int8_t *binary, unsigned bin_size);
 void print_instruction(unsigned addr, instruction_t inst, unsigned inst_size, u_int8_t *binary, bool space);
-
+bool has_reg(const instruction_t *inst);

src/instructions.c

@@ -1,6 +1,7 @@
 #include <stddef.h>
 #include <stdio.h>
 #include "dasm.h"
+#include "instructions.h"
 
 instruction_t parse_inst(u_int8_t *binary, unsigned long size)
 {
@@ -27,8 +28,8 @@ const instruction_t instructions[] = {
 	{.opcode = 0x01, .extended = -1, .name = "add", .mode = {R_M16, REG16, END}, .exec = NULL},
 	{.opcode = 0x02, .extended = -1, .name = "add", .mode = {REG8, R_M8, END}, .exec = NULL},
 	{.opcode = 0x03, .extended = -1, .name = "add", .mode = {REG16, R_M16, END}, .exec = NULL},
-	{.opcode = 0x04, .extended = -1, .name = "add al", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0x04, .extended = -1, .name = "add al", .mode = {OPREG8, IMM8, END}, .exec = NULL},
-	{.opcode = 0x05, .extended = -1, .name = "add ax", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0x05, .extended = -1, .name = "add ax", .mode = {OPREG16, IMM16, END}, .exec = NULL},
 	{.opcode = 0x06, .extended = -1, .name = "push es", .mode = {END}, .exec = NULL},
 	{.opcode = 0x07, .extended = -1, .name = "pop es", .mode = {END}, .exec = NULL},
 	{.opcode = 0x08, .extended = -1, .name = "or", .mode = {R_M8, REG8, END}, .exec = NULL},
@@ -143,13 +144,13 @@ const instruction_t instructions[] = {
 	{.opcode = 0x85, .extended = -1, .name = "test", .mode = {R_M16, REG16, END}, .exec = NULL},
 	{.opcode = 0x86, .extended = -1, .name = "xchg", .mode = {R_M8, REG8, END}, .exec = NULL},
 	{.opcode = 0x87, .extended = -1, .name = "xchg", .mode = {R_M16, REG16, END}, .exec = NULL},
-	{.opcode = 0x88, .extended = -1, .name = "mov", .mode = {R_M8, REG8, END}, .exec = NULL},
+	{.opcode = 0x88, .extended = -1, .name = "mov", .mode = {R_M8, REG8, END}, .exec = &mov},
-	{.opcode = 0x89, .extended = -1, .name = "mov", .mode = {R_M16, REG16, END}, .exec = NULL},
+	{.opcode = 0x89, .extended = -1, .name = "mov", .mode = {R_M16, REG16, END}, .exec = &mov},
-	{.opcode = 0x8A, .extended = -1, .name = "mov", .mode = {REG8, R_M8, END}, .exec = NULL},
+	{.opcode = 0x8A, .extended = -1, .name = "mov", .mode = {REG8, R_M8, END}, .exec = &mov},
-	{.opcode = 0x8B, .extended = -1, .name = "mov", .mode = {REG16, R_M16, END}, .exec = NULL},
+	{.opcode = 0x8B, .extended = -1, .name = "mov", .mode = {REG16, R_M16, END}, .exec = &mov},
-	// {.opcode = 0x8C, .extended = -1, .name = "mov", .mode = {R_M16, sreg, END}, .exec = NULL},
+	// {.opcode = 0x8C, .extended = -1, .name = "mov", .mode = {R_M16, sreg, END}, .exec = &mov},
 	{.opcode = 0x8D, .extended = -1, .name = "lea", .mode = {REG16, R_M16, END}, .exec = NULL},
-	// {.opcode = 0x8E, .extended = -1, .name = "mov", .mode = {sreg, R_M16, END}, .exec = NULL},
+	// {.opcode = 0x8E, .extended = -1, .name = "mov", .mode = {sreg, R_M16, END}, .exec = &mov},
 	{.opcode = 0x8F, .extended = -1, .name = "pop", .mode = {R_M16, END}, .exec = NULL},
 	{.opcode = 0x90, .extended = -1, .name = "xchg ax, ax", .mode = {END}, .exec = NULL},
 	{.opcode = 0x91, .extended = -1, .name = "xchg cx, ax", .mode = {END}, .exec = NULL},
@@ -167,10 +168,10 @@ const instruction_t instructions[] = {
 	{.opcode = 0x9D, .extended = -1, .name = "popf", .mode = {END}, .exec = NULL},
 	{.opcode = 0x9E, .extended = -1, .name = "sahf", .mode = {END}, .exec = NULL},
 	{.opcode = 0x9F, .extended = -1, .name = "lahf", .mode = {END}, .exec = NULL},
-	// {.opcode = 0xA0, .extended = -1, .name = "mov al", .mode = {moffs16, END}, .exec = NULL},
+	// {.opcode = 0xA0, .extended = -1, .name = "mov al", .mode = {moffs16, END}, .exec = &mov},
-	// {.opcode = 0xA1, .extended = -1, .name = "mov ax", .mode = {moffs16, END}, .exec = NULL},
+	// {.opcode = 0xA1, .extended = -1, .name = "mov ax", .mode = {moffs16, END}, .exec = &mov},
-	// {.opcode = 0xA2, .extended = -1, .name = "mov %s, al", .mode = {moffs16, END}, .exec = NULL},
+	// {.opcode = 0xA2, .extended = -1, .name = "mov %s, al", .mode = {moffs16, END}, .exec = &mov},
-	// {.opcode = 0xA3, .extended = -1, .name = "mov %s, ax", .mode = {moffs16, END}, .exec = NULL},
+	// {.opcode = 0xA3, .extended = -1, .name = "mov %s, ax", .mode = {moffs16, END}, .exec = &mov},
 	{.opcode = 0xA4, .extended = -1, .name = "movsb", .mode = {END}, .exec = NULL},
 	{.opcode = 0xA5, .extended = -1, .name = "movsw", .mode = {END}, .exec = NULL},
 	{.opcode = 0xA6, .extended = -1, .name = "cmpsb", .mode = {END}, .exec = NULL},
@@ -183,29 +184,29 @@ const instruction_t instructions[] = {
 	{.opcode = 0xAD, .extended = -1, .name = "lodsw", .mode = {END}, .exec = NULL},
 	{.opcode = 0xAE, .extended = -1, .name = "scasb", .mode = {END}, .exec = NULL},
 	{.opcode = 0xAF, .extended = -1, .name = "scasw", .mode = {END}, .exec = NULL},
-	{.opcode = 0xB0, .extended = -1, .name = "mov al", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB0, .extended = -1, .name = "mov al", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB1, .extended = -1, .name = "mov cl", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB1, .extended = -1, .name = "mov cl", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB2, .extended = -1, .name = "mov dl", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB2, .extended = -1, .name = "mov dl", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB3, .extended = -1, .name = "mov bl", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB3, .extended = -1, .name = "mov bl", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB4, .extended = -1, .name = "mov ah", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB4, .extended = -1, .name = "mov ah", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB5, .extended = -1, .name = "mov ch", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB5, .extended = -1, .name = "mov ch", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB6, .extended = -1, .name = "mov dh", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB6, .extended = -1, .name = "mov dh", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB7, .extended = -1, .name = "mov bh", .mode = {IMM8, END}, .exec = NULL},
+	{.opcode = 0xB7, .extended = -1, .name = "mov bh", .mode = {IMM8, END}, .exec = &mov},
-	{.opcode = 0xB8, .extended = -1, .name = "mov ax", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xB8, .extended = -1, .name = "mov ax", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xB9, .extended = -1, .name = "mov cx", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xB9, .extended = -1, .name = "mov cx", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBA, .extended = -1, .name = "mov dx", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBA, .extended = -1, .name = "mov dx", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBB, .extended = -1, .name = "mov bx", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBB, .extended = -1, .name = "mov bx", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBC, .extended = -1, .name = "mov sp", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBC, .extended = -1, .name = "mov sp", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBD, .extended = -1, .name = "mov bp", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBD, .extended = -1, .name = "mov bp", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBE, .extended = -1, .name = "mov si", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBE, .extended = -1, .name = "mov si", .mode = {IMM16, END}, .exec = &mov},
-	{.opcode = 0xBF, .extended = -1, .name = "mov di", .mode = {IMM16, END}, .exec = NULL},
+	{.opcode = 0xBF, .extended = -1, .name = "mov di", .mode = {IMM16, END}, .exec = &mov},
 
 	{.opcode = 0xC2, .extended = -1, .name = "ret", .mode = {IMM16, END}, .exec = NULL},
 	{.opcode = 0xC3, .extended = -1, .name = "ret", .mode = {END}, .exec = NULL},
 	{.opcode = 0xC4, .extended = -1, .name = "les", .mode = {REG16, R_M16, END}, .exec = NULL},
 	{.opcode = 0xC5, .extended = -1, .name = "lds", .mode = {R_M16, REG16, END}, .exec = NULL},
-	{.opcode = 0xC6, .extended = -1, .name = "mov", .mode = {R_M16, IMM8, END}, .exec = NULL},
+	{.opcode = 0xC6, .extended = -1, .name = "mov", .mode = {R_M16, IMM8, END}, .exec = &mov},
-	{.opcode = 0xC7, .extended = -1, .name = "mov", .mode = {R_M16, IMM16, END}, .exec = NULL},
+	{.opcode = 0xC7, .extended = -1, .name = "mov", .mode = {R_M16, IMM16, END}, .exec = &mov},
 	{.opcode = 0xCA, .extended = -1, .name = "retf", .mode = {IMM16, END}, .exec = NULL},
 	{.opcode = 0xCB, .extended = -1, .name = "retf", .mode = {END}, .exec = NULL},
 	{.opcode = 0xCC, .extended = -1, .name = "int3", .mode = {END}, .exec = NULL},

src/instructions.h

@@ -0,0 +1,5 @@
+#pragma once
+
+#include "dasm.h"
+
+void mov(const instruction_t *self, state_t *state);

src/instructions/memory.c

@@ -0,0 +1,98 @@
+#include <stdint.h>
+#include <stddef.h>
+#include <stdio.h>
+#include "../dasm.h"
+
+unsigned *get_reg_operand(state_t *state, bool is16bit)
+{
+	uint8_t *registers8[8] = {
+		&state->al,
+		&state->cl,
+		&state->dl,
+		&state->bl,
+		&state->ah,
+		&state->ch,
+		&state->dh,
+		&state->bh,
+	};
+	uint16_t *registers16[8] = {
+		&state->ax,
+		&state->cx,
+		&state->dx,
+		&state->bx,
+		&state->sp,
+		&state->bp,
+		&state->si,
+		&state->di,
+	};
+
+	if (is16bit)
+		return (void *)registers16[(state->binary[state->pc + 1] & 0b111000) >> 3];
+	return (void *)registers8[(state->binary[state->pc + 1] & 0b111000) >> 3];
+}
+
+unsigned *get_rm_operand(const instruction_t *inst, state_t *state, unsigned *imm_idx, bool is16bit)
+{
+	unsigned mod = state->binary[1] >> 6;
+	unsigned rm = state->binary[1] & 0b111;
+
+	if (mod == 0b11)
+		return get_reg_operand(state, is16bit);
+
+	if (mod == 0 && rm == 0b110) {
+		*imm_idx += 2;
+		return (void *)&state->binary[state->pc + *imm_idx - 2];
+	}
+	// TODO:
+	return NULL;
+}
+
+unsigned *get_operand(const instruction_t *inst, unsigned i, state_t *state)
+{
+	unsigned *ret = NULL;
+	unsigned imm_idx = 0;
+
+	switch (inst->mode[i]) {
+	case IMM16:
+		imm_idx++;
+		FALLTHROUGHT;
+	case IMM8:
+		imm_idx++;
+		ret = (void *)&state->binary[state->pc + state->parse_data.imm_idx];
+		break;
+	case REL16:
+		imm_idx++;
+		FALLTHROUGHT;
+	case REL8:
+		imm_idx++;
+		ret = NULL;
+		// TODO:
+		break;
+	case REG8:
+		ret = get_reg_operand(state, false);
+		break;
+	case REG16:
+		ret = get_reg_operand(state, true);
+		break;
+	case R_M8:
+		ret = get_rm_operand(inst, state, &imm_idx, false);
+		break;
+	case R_M16:
+		ret = get_rm_operand(inst, state, &imm_idx, true);
+		break;
+	case END:
+		printf("Invalid parameter read.");
+		break;
+	}
+	state->parse_data.imm_idx += imm_idx;
+	return ret;
+}
+
+void mov(const instruction_t *self, state_t *state)
+{
+	unsigned *from = get_operand(self, 0, state);
+	unsigned *to = get_operand(self, 1, state);
+
+	printf("mov %p %p\n", from, to);
+	printf("mov %x %x\n", (uint16_t)*from, (uint16_t)*to);
+}

src/interpretor.c

@@ -37,6 +37,7 @@ int interpret(u_int8_t *binary, unsigned long size)
 	}
 
 	binary += header_size;
+	state->binary = binary;
 
 	printf(" AX   BX   CX   DX   SP   BP   SI   DI  FLAGS IP\n");
 	while (state->pc < size) {
@@ -47,8 +48,9 @@ int interpret(u_int8_t *binary, unsigned long size)
 		}
 		print_state(state);
 		print_instruction(state->pc, inst, inst_size, binary, false);
+		state->parse_data.imm_idx = 1 + (inst.extended != -1 || has_reg(&inst));
 		if (inst.exec)
-			inst.exec(state);
+			inst.exec(&inst, state);
 		else
 			printf("Not implemented.\n");
 		state->pc += inst_size;

src/interpretor.h

@@ -53,4 +53,10 @@ typedef struct state {
 		};
 		unsigned flags:10;
 	};
+
+	uint8_t *binary;
+
+	struct {
+		unsigned imm_idx;
+	} parse_data;
 } state_t;