The Iris project was born just like many other ambitious undertakings: by looking at a huge, decades-old system developed and maintained by thousands of professionals (the Linux kernel) and going:
"Eh, I could do that in a few weeks. How hard could it possibly be?"
Oh, Past Connor...
Very, very difficult.
Future Connor is now learning that the hard way.
Is there a special reason you named the project that?
Is it named after someone you love? Does it have a connection to the project?
Nope.
It literally just sounded cool, and all the other good names were taken.
(And I like the blue-purple color that irises sometimes are. Bonus points!)
Actually, yes! Wait, how'd you know that...
When I was beginning this project, there were two huge differences: the kernel was i386 (still x86, but 32-bit), and the entire operating system was being built at the same time as the kernel. I was planning on this being an easy peasy, month-long-ish project, so I named it TLOS (pronounced "tee-loss"), which stands for "Tiny Little Operating System".
Then, I decided that I needed a separate name for the kernel! So, given the central nature of kernels, I chose the name "Core Kernel", which was then shortened to "CorK".
Since I love uniform naming schemes, I then renamed the entire project to Vial.
Why, you ask?
Because vials are experimental equipment, sure...
But also...
...you put CorKs in Vials.
Just give Past Connor a courtesy laugh and keep reading.
I know, it's awful.
After the project gained steam, I decided that CorK and Vial were... relatively childish-sounding names for a project that will probably last multiple years. So, after going through tons of ideas (and learning that all of the good ones were already taken), I settled on Iris!
I love the color that some irises are (which will become the accent color for the future UI), and it's just a really nice name in general anyway. Short, punchy, easy to remember, and most importantly: not taken!
The future operating system built around Iris will be named Lavender. I told you already: I love uniform naming schemes! Same color palette, same vibe, while also sounding more flowing.
It all just felt perfect.
Okay, okay, fine.
Iris is an x86_64 OS kernel written mostly in C, with an assembly stub for entry to set up things like the stack.
Iris boots via its custom UEFI bootloader (CUB v2.1), which passes a BootInfo struct to the kernel. After CUB exits UEFI boot services, Iris takes control of the machine and no longer relies on UEFI or CUB.
Iris currently uses the QEMU debug port at 0xE9 to print to a terminal as its main source of communication with the user. In the future, a framebuffer text printer will be developed and replace the debug port.
On the developer's side, Iris is built using a Makefile that compiles using GCC and LD, producing a kernel ELF file and CUB EFI executable. You can use the Dockerfile in the repository to create a container and run make inside of it (instructions are at the bottom of this README).
As always, this project evolves quickly, so it's best to check the source files yourself instead of relying on my poorly-made descriptions here!
In Docker, run these commands:
-
To build the Docker image, move into the root of this repository and run:
-
On x84_64 systems (Linux)
docker build -t iris-i . -
If you are running Docker on an ARM machine (including a Mac):
docker build --platform linux/amd64 -t iris-i .
-
-
Then, to spin up a temporary container that will delete itself upon exit:
docker run --rm -it --name iris-c --user $(id -u):$(id -g) -v "$(pwd)":/home/iris iris-i
(Remove --rm to create a persistent container instead)
-
Then, inside the container, run:
make clean && make -
Now, choose your own adventure!
-
If you have QEMU installed on the host computer (
qemu-system-x86), on the host run:make run -
If not, no worries! Inside of the container, you can use:
make run-c
-