sudo apt update && sudo apt install nasm

For MSYS2, Arch Linux or Linux distro that uses pacman-package-manager (Arch-based):

Update pacman, then install nasm compiler:

sudo pacman -Syu && sudo pacman --sync nasm

For Red Hat Linux or the Linux distro that uses yum package manager (RPM-based):

Update yum, then install nasm compiler:

sudo yum -y update && sudo yum install nasm

Check off whether NASM is downloaded:

nasm --version    #     It should output "NASM version <version>"
ndisasm -v        #  -- It should output "NDISASM version <version>"

Set up a new Intergrated Develop Environment in order to take down Netwide Assembler script easier

Please see my contents about Visual Studio Code, CoC and Neovim

Create your first Netwide Assembler source code

We will use Neovim for example to open a file. However, you can also open it using Vim or VS Code:

nvim main.nasm

Then type: (This step needs who5673-nasm or coc-nasm)

printHello

or ph, pH,...anything that relates to snippet printHello.

nvim-cmp will appear the snippet printHello for you.
It should output this script after choosing that snippet:

%macro exit 1
  mov rax, 60
  mov rdi, %1
  syscall
%endmacro

%macro getstr 2
  mov rax, 1
  mov rdi, 1
  mov rsi, %1
  mov rdx, %2
  syscall
%endmacro

section .data
  example db 0x1b, "[1;38;2;255;165;0m", "Hello world", 0x1b, "[0m", 0xA
  length equ $-example

section .text
  global _start

_start:
  getstr example, length
  exit 0

Then use nasm to compile your Netwide Assembler file:

nasm -f elf64 main.nasm -o main.o    # Compile NASM file into an object file using specified format
ld main.o -o main   # Link the object file to the binary, then make a binary file

Attention: You need to choose right format type in order to make the binary work.

According to nasm -hf instructions, we have some format like this:

• win32 (legacy alias: win): Windows 32-bit
• win64: Windows 64-bit
• macho32 (legacy alias: macho): Mach-O i386 MacOS X and variants (32-bit)
• macho64: Mach-O x86_64 MacOS X and variants (64-bit)
• elf32 (legacy alias: elf): x86 Linux distros (32-bit)
• elf64: x86-64 Linux distros (64-bit)
• bin: flat RAW binary file.
  ⚠️ You should not use over 16-bit Netwide Assembler syntax for this type as BIOS real-mode cannot load greater than or equal to 32-bit registers.

Disassemble a RAW binary file (.bin, .hex) using ndisasm

Create a raw binary file. For instance: (ATTENTION: THIS IS THE COMMAND FOR x86-64 PROCESSOR MODE)

echo -ne '\x48\xb8\x01\x00\x00\x00\x00\x00\x00\x00\x48\xbf\x01\x00\x00\x00\x00\x00\x00\x00\x48\xbe\x00\x10\x60\x00\x00\x00\x00\x00\x48\xba\x0c\x00\x00\x00\x00\x00\x00\x00\x0f\x05\x48\xb8\x3c\x00\x00\x00\x00\x00\x00\x00\x48\x31\xff\x0f\x05' > main.bin

Disassemble a RAW binary file (not real binary files like rm, cp, ls,...):

ndisasm -b 64 main.bin    # You can use -b 16, 32, 64 if you want to use other preprocessors

If you use 64-bit preprocessor, it should output those texts like this (can be different from different processor modes).

ndisasm's Disassembly result

00000000  48B8010000000000  mov rax,0x1
         -0000
0000000A  48BF010000000000  mov rdi,0x1
         -0000
00000014  48BE001060000000  mov rsi,0x601000
         -0000
0000001E  48BA0C0000000000  mov rdx,0xc
         -0000
00000028  0F05              syscall
0000002A  48B83C0000000000  mov rax,0x3c
         -0000
00000034  4831FF            xor rdi,rdi
00000037  0F05              syscall

Get started with Netwide Assembler

Not like mid-level and high-level programming languages, NASM is a low-level one and it is much harder to find documents for learning it. Moreover, before programming this language, you need to know some skills about basic things to deal with the CPU and compiler.

For more information about the nasm compiler, please go to the real document page.

  int main(void) {
  return 0;
  }
          

1. Hexdecimal number and address

Hexdecimal is really important to deal with the data and initialize a value into things like:

• Register: Things in the CPU that can be writen or overwritten a value like hexdecimal integers in order to make an interrupt like you want.

Copy

; I will tell you how to use the `mov` command later
mov ax, 0x5307    ; (16-bit) APM Basic Input/Output System (BIOS) function: Set power state
mov bx, 0x0001    ; (16-bit) Device ID: All devices
mov cx, 0x0003    ; (16-bit) Power State: Off
int 0x15          ; (16-bit) APM Interrupt
      

• Memory address: Used as an address to save an array of ASCII characters. This is the only method that helps you print strings.
  
  Memory syntax example in NASM with default abs:
  • [0x601000] :absolute;
  • [rel msg]: relative label example address named msg;
  • [msg]: Label memory address that is defined
  • NOTE: Some registers like rsi does not need square brackets to make it be an address. For instance:

  Copy

  ; I will tell you how to use the `mov`, `section .data` and `db` commands later
  mov rsi, msg
  
  section .data
    msg db "Hello world", 0xa, 0x0
              

• Immediate: An array of ASCII characters that is saved into a label which is an alias of something like integer value and memory address.

Copy

msg db "Example string", 0xa, 0x0    ; db: Define byte a character or an array of characters, each character takes 1 byte (8-bit)
len equ $-msg                        ; Calculates the length of the array by subtracting current address by msg label offset
                                     ; equ: Define a constant