/brainfuck-in-the-bootsector

Primary LanguageAssemblyMIT LicenseMIT

Brainfuck in the Boot Sector

This is a (mostly) functional BF interpreter that's run entirely on the boot sector.

Bootstrapped off of https://github.com/belamenso/nasm-bfi by belamenso.

See boot.asm for implementation details.

Build and Run

$ nasm -f bin boot.asm -o boot.bin
$ qemu-system-x86_64 boot.bin  # or qemu-system-i686

Alt+2, then 'q' then 'ENTER' to exit QEMU.

Usage

Type or copy and paste BF code into the window, then hit ENTER to interpret.

Try copying this Hello World example:

++++++++[>++++[>++>+++>+++>+<<<<-]>+>+>->>+[<]<-]>>.>---.+++++++..+++.>>.<-.<.+++.------.--------.>>+.>++.

If there is anything that should be printed, it will appear next to the Output> at the top of the prompt.

After execution, the instructions will still be on the prompt, but you can overwrite the instructions.

Also after execution, the state of the tape will not reset back to how it was at startup, so you can constantly change it with every interpretter run.

  • As a result of this, it could technically be possible to get stuck in an infinite loop of you hit ENTER a couple of times on the same Hello world code.

Booting on Real Hardware

Take a USB stick, burn the binary on it, and select the USB stick from the BIOS when booting.

Assuming your USB stick is mounted at /dev/sdb, just run:

$ sudo dd if=boot.bin of=/dev/sdb && sync

Constraints

A lot of corners had to be cut obviously for the sake of fitting it all into 512 bytes. Here are some of them:

  • The tape that the BF code is only (at least) 64 bytes (instead of the 30000 recommended on the wiki). This could technically be expanded though as the remainder of the kernel if I kept reading from more boot sectors ;)
  • Instead of also dedicating a section in the bootsector for storing BF code to be interpretted, I opted for instead just reading them off the screen using the various int 0x10 interupt functions.
  • I have not registered arrow key inputs for the terminal.
  • I have not implemented any sort of checks for balanced [s and ]s.
  • To avoid printing characters that could set the cursor past the Output line in the terminal, I only print the original character if ' ' <= char <= '~'. This avoids printing stuff like newlines or carriage returns.