First assigment done (w/o writing to file)

1-1/compile.sh

@@ -0,0 +1,3 @@
+#! /usr/bin/env bash
+
+gcc 1-1/decoder.c -O0 -g -gdwarf -o 1-1/8086coded

1-1/decoder.c

@@ -0,0 +1,104 @@
+#include <stdint.h>
+#include <stdio.h>
+#include <limits.h>
+
+enum DT_INSTRUCTIONS
+	{
+		MOV_RM_TF_R = 0b1000'10, //d w
+		MOV_I_T_RM  = 0b1100'011, // w
+		MOV_I_T_R   = 0b1011, // w reg
+		MOV_M_T_A   = 0b1010'000, // w
+		MOV_A_T_M   = 0b1010'001, // w
+		MOV_RM_T_SR = 0b1000'1110, 
+		MOV_SR_T_RM = 0b1000'1100, 
+	};
+
+enum MOV_REGISTER_MODE
+	{
+		MEM_NO_DISP    = 0b00,
+		MEM_8BIT_DISP  = 0b01,
+		MEM_16BIT_DISP = 0b010,
+		REG_NO_DISP    = 0b11
+	};
+
+
+char REG_ENCODING_TXT[0b10000][3] = { "al", "cl", "dl", "bl", "ah", "ch", "dh", "bh",
+                                   "ax", "cx", "dx", "bx", "sp", "bp", "si", "di"};
+enum REG_ENCODING
+	{
+		R_AL = 0b000,
+		R_CL = 0b001,
+		R_DL = 0b010,
+		R_BL = 0b011,
+		R_AH = 0b100,
+		R_CH = 0b101,
+		R_DH = 0b110,
+		R_BH = 0b111
+	};
+
+enum WREG_ENCODING
+	{
+		WR_AX = 0b000,
+		WR_CX = 0b001,
+		WR_DX = 0b010,
+		WR_BX = 0b011,
+		WR_SP = 0b100,
+		WR_BP = 0b101,
+		WR_SI = 0b110,
+		WR_DI = 0b111
+	};
+
+typedef struct
+{
+	FILE *binary;
+	uint64_t size;	
+} binary_data;
+
+
+void MOV_RM_TF_R_parse(uint8_t byte1, uint8_t byte2, uint8_t byte3, uint8_t byte4)
+{
+	const uint8_t wide_mask = 0b0000'0001;
+	const uint8_t dest_mask = 0b0000'0010;
+	const uint8_t mod_mask  = 0b1100'0000; 
+	const uint8_t regm_mask = 0b0011'1000;
+	
+    if ((byte2 & mod_mask) != mod_mask) return;
+	//Since we're only doing register mode/register to register, both reg and r/m are affected by the W bit
+	bool reg_is_dest = (byte1 & dest_mask);
+	//if D=1, dest is retrieved from the reg field in byte2 and src from r/m
+	uint8_t dest_value = (reg_is_dest) ? ((byte2 & regm_mask) >> 3) + ((byte1 & wide_mask) * 8) : (byte2 & (regm_mask >> 3)) + ((byte1 & wide_mask) * 8);
+	uint8_t src_value = (reg_is_dest) ? (byte2 & (regm_mask >> 3)) + ((byte1 & wide_mask) * 8) : ((byte2 & regm_mask) >> 3) + ((byte1 & wide_mask) * 8);
+    printf("%s, %s\n", REG_ENCODING_TXT[dest_value], REG_ENCODING_TXT[src_value]);
+	return;
+}
+
+int main(int argc, char** argv)
+{
+	if (argc != 2) return -1;
+	if (CHAR_BIT != 8) return -2;
+
+	binary_data bin;
+	unsigned char byte[1];
+	bin.binary = fopen(argv[1], "rb");
+	fseek(bin.binary, 0, SEEK_END);
+	bin.size = ftell(bin.binary);
+	fseek(bin.binary, 0, SEEK_SET);
+
+	for (int bytes_read = 0; bytes_read < bin.size; bytes_read++)
+	{
+		fread(byte, sizeof(byte), 1, bin.binary);
+		uint8_t manip_inst = (uint8_t)byte[0];
+		uint8_t inst = (uint8_t)byte[0];
+		//For now we're just checking for RM_TF_R
+		manip_inst = (manip_inst >> 2);
+		if (manip_inst != MOV_RM_TF_R) break;
+		printf("%s ", "mov");
+		if (bytes_read >= bin.size) break;
+    	fread(byte, sizeof(byte), 1, bin.binary);
+		bytes_read++;
+		uint8_t inst_byte2 = (uint8_t)byte[0];
+    	MOV_RM_TF_R_parse(inst, inst_byte2, 0, 0);
+	}
+
+	return 0;
+}

1-1/listing-37

@@ -0,0 +1 @@
+菓

1-1/listing-37.asm

@@ -0,0 +1,3 @@
+bits 16
+	
+mov cx, bx

1-1/listing-38

@@ -0,0 +1 @@
+‰Ùˆå‰Ú‰Þ‰ûˆÈˆí‰Ã‰ó‰ü‰Å

1-1/listing-38.asm

@@ -0,0 +1,13 @@
+bits 16
+
+mov cx, bx
+mov ch, ah
+mov dx, bx
+mov si, bx
+mov bx, di
+mov al, cl
+mov ch, ch
+mov bx, ax
+mov bx, si
+mov sp, di
+mov bp, ax

1-1/nasm-compile.sh

@@ -0,0 +1,4 @@
+#! /usr/env/bin bash
+
+nasm -f bin 1-1/listing-38.asm -o 1-1/listing-37
+nasm -f bin 1-1/listing-38.asm -o 1-1/listing-38