I am writing my own OS

5 minute read Published:

Kernel panic!

My fascination with computers turned faster as an obsession.

I’m obsessed with my machine.
Really, I’m not feeling all right if I don’t clean up my machine physically every week, and if I don’t reinstall a new clean operating system each 6 months.
Obviously, before each clean install, I overwrite my entire hard drive with zeros… which takes me a full day for a 1 TB hard drive.

When I was using a GNU/Linux OS, and before installing a new operating system on my computer, I always look up on distrowatch what’s new in the GNU/Linux world.
What is the new fashion OS, new version of my favourite OSs, new forks, new OSs, etc…
So, it happened (a lot) that I install a new operating system each 6 months, just to test it, see what’s new, the stability… and reinstall an ArchLinux maybe 3 days after.

Archlinux, Debian and ElementaryOS are my favourite OSs.
Archlinux because it’s really modulable, and you can install what you want without have some garbage applications like basic mail clients, video and music players, or even Amazon packets (just 4 you Ubuntu…).
Debian is really stable, but I install it only if my computer have more than 2 years, just to be sure that packets to manage for example my GPU is accessible and can be install/run without make some wizard things.
Finally, I use ElementaryOS a lot for recent machines, especially because the base of this OS is Ubuntu, and I didn’t have some troubles using Ubuntu packets for recent hardware. ElementaryOS is pretty stable and Pantheon, the desktop, is really ergonomical and beautiful - inspired by macOS.

The operating system is the second software that the machine launch, after the bios.
The operating system manage, by itself, some sort of things: processes, interruption handlers, memory management, events from external devices like Keyboard, etc… It’s a magnificient piece of engineering, between you and the hardware.

BUT, I don’t know how works really an operating system. I mean, I take some courses like “Architecture of Operating Systems” at university, I’ve got a large comprehension of what an OS do and how he can do that, but I never programed one. And, in computer science, like some teachers and research guys told me: “If you want to learn or just want to know if you had understand something, try to program it!”

I’m using Rust since 1 year now, and it’s a wonderful programming language. This language has been developed by the Mozilla foundation to avoid some problems due to using C++ for their products, especially memory leaks and error management. In june, a first version of Servo, the faster and most memery efficient modern browser engine ever made, has been released in a basic web browser - just a public ALPHA. And the project is REALLY AWE-SOME!
Rust rised up recently the 1.10 stable version, and a lot of awesome projects are currently been build using this language.

Personaly, at the beginning, I used Rust a lot to port my slow Python projects into a faster and memory efficient program, but also during my combinatorial optimization courses, to prone usage of modern programming languages and also “4 fun”.
During October/November 2015, I started to contribute to RedoxOS, a hobby operating system. RedoxOS works actually on Qemu and VirtualBox, and is written in Rust. The perfect project for me to contribute to a cool project, learn some things about operating systems and to improve myself in a cool programming language like Rust.

But the project is really big, there is a lot of contributions per week and the code is optimized to be not understandable sometimes…

So, to perform myself in 1) contributing to build the best operating system in Rust as possible and 2) improving myself in Rust, I decided to create my own operating system in Rust.

I called it arOS, for another rust Operating System.

Steve Klabnik, a famous Rubyist, just started a project called intermezzOS. This project is an hobby operating system, written in Rust. intermezzOS is open source, and anyone can see the code and reflexions behind this project, using the online book.
Steve made this project by himself, “just 4 fun”. Also, as the project is currently building, he release some news and code regularly.

Currently, the final chapter is about how to link Rust code in your boot assembly code, which is the base to begin to program your own OS in Rust.

During 4 days, I followed the tutorial of Steve Klabnik. My code is kinda similar to his one, but I’m completing my learning programming a VGA driver by myself, which is the next chapter of the tutorial (and it’s not easy, trust me…).

During these 4 days, I was fighting with the nightly version of Rust and his feature tags. These tags are incredibly usefull, but there is much features for the nightly version that I’m completely lost sometimes… It was the first time that I creating some extern and unsafe code in Rust, that is really strange and completely exciting!
Also, I wrote some assembly code during my classes, but not as well as the boot and multiboot assembly code showed in the tutorial.

This project is really interesting, and I just learn usefull things about assembly code, how to link Rust code, and some optimizations made by rustc, the Rust compiler, for projects which has to use less memory as possible.

This project is not really serious, just for learn and just for fun :-)

My next step, after the VGA driver, is to port this OS for the ARM family (a Raspberry Pi for example), using Qemu to test all long the operating system.