The jsix operating system
jsix is a custom multi-core x64 operating system that I am building from scratch. It's far from finished, or even being usable (see the Status and Roadmap section, below) but all currently-planned major kernel features are now implemented to at least a passable level.
The design goals of the project are:
Modernity - I'm not interested in designing for legacy systems, or running on all hardware out there. My target is only 64 bit architectures, and modern commodity hardware. Currently that means x64 systems with Nehalem or newer CPUs and UEFI firmware. (See this list for the currently required CPU features.) Eventually I'd like to work on an AArch64 port, partly to force myself to factor out the architecture-dependent pieces of the code base.
Modularity - I'd like to pull as much of the system out into separate processes as possible, in the microkernel fashion. A sub-goal of this is to explore where the bottlenecks of such a microkernel are now, and whether eschewing legacy hardware will let me design a system that's less bogged down by the traditional microkernel problems.
Exploration - I'm really mostly doing this to have fun learning and exploring modern OS development. Initial feature implementations may temporarily throw away modular design to allow for exploration of the related hardware.
A note on the name: This kernel was originally named Popcorn, but I have since
discovered that the Popcorn Linux project is also developing a kernel with that
name, started around the same time as this project. So I've renamed this kernel
jsix (Always styled jsix or
j6, never capitalized) as an homage to L4, xv6,
and my wonderful wife.
Status and Roadmap
The following major feature areas are targets for jsix development:
UEFI boot loader
Done. The bootloader loads the kernel and initial userspace programs, and sets up necessary kernel arguments about the memory map and EFI GOP framebuffer. Possible future ideas:
- take over more init-time functions from the kernel
- rewrite it in Zig
Virtual memory: Sufficient. The kernel manages virtual memory with a number
of kinds of
vm_area objects representing mapped areas, which can belong to
one or more
vm_space objects which represent a whole virtual memory space.
(Each process has a
vm_space, and so does the kernel itself.)
Remaining to do:
- TLB shootdowns
- Page swapping
- Large / huge page support
Physical page allocation: Sufficient. The current physical page allocator implementation uses a group of blocks representing up-to-1GiB areas of usable memory as defined by the bootloader. Each block has a three-level bitmap denoting free/used pages.
- Align blocks so that their first page is 1GiB-aligned, making finding free large/huge pages easier.
Sufficient. The global scheduler object keeps separate ready/blocked lists per core. Cores periodically attempt to balance load via work stealing.
User-space tasks are able to create threads as well as other processes.
Syscalls: Sufficient. User-space tasks are able to make syscalls to the
kernel via fast SYSCALL/SYSRET instructions. Syscalls made via
libj6 look to
both the callee and the caller like standard SysV ABI function calls. The
implementations are wrapped in generated wrapper functions which validate the
request, check capabilities, and find the appropriate kernel objects or handles
before calling the implementation functions.
IPC: Working, needs optimization. The current IPC primitives are:
- Mailboxes: endpoints for asynchronously-delivered small messages. Currently these messages are double-copied - once from the caller into a kernel buffer, and once from the kernel to the receiver. This works and is not a major cause of slowdown, but will need to be optimized in the future.
- Channels: endpoints for asynchronous uni-directional streams of bytes. Currently these also suffer from a double-copy problem, and should probably be replaced eventually by userspace shared memory communication.
- Events: objects that can be signalled to send asynchronous notifications to waiting threads.
- Framebuffer driver: In progress. Currently on machines with a video device accessible by UEFI, jsix starts a user-space framebuffer driver that only prints out kernel logs.
- Serial driver: In progress. The current UART currently only exposes COM1 as an output channel, but no input or other serial ports are exposed.
- USB driver: To do
- AHCI (SATA) driver: To do
jsix uses the Ninja build tool, and generates the build files for it with
configure script. The build also relies on a custom toolchain sysroot, which can be
downloaded or built using the scripts in jsix-os/toolchain.
Other build dependencies:
configure script has some Python dependencies - these can be installed via
pip, though doing so in a python virtual environment is recommended.
pip will also install
A Debian 11 (Bullseye) system can be configured with the necessary build dependencies by running the following commands from the jsix repository root:
sudo apt install clang lld nasm mtools python3-pip python3-venv python3 -m venv ./venv source venv/bin/activate pip install -r requirements.txt peru sync
Setting up the sysroot
Build or download the toolchain sysroot as mentioned above with
jsix-os/toolchain, and symlink the built toolchain directory as
at the root of this project.
# Example if both the toolchain and this project are cloned under ~/src ln -s ~/src/toolchain/toolchains/llvm-13 ~/src/jsix/sysroot
Building and running jsix
Once the toolchain has been set up, running the
./configure script (see
./configure --help for available options) will set up the build configuration,
ninja -C build (or wherever you put the build directory) will actually run
the build. If you have
qemu-system-x86_64 installed, the
qemu.sh script will
to run jsix in QEMU
I personally run this either from a real debian amd64 bullseye machine or a windows WSL debian bullseye installation. Your mileage may vary with other setups and distros.
Running the test suite
jsix now has the
test_runner userspace program that runs various automated
tests. It is not included in the default build, but if you use the
manifest it will be built, and can be run with the
test.sh script or the
./configure --manifest=assets/manifests/test.yml if ./test.sh; then echo "All tests passed!"; else echo "Failed."; fi