umbongo

An x86 operating system

• 64-bit and UEFI only
• Aiming for a microkernel design, with the kernel's primary focus being IPC
• Work in progress posix layer library that can already host a few ports (ld, tcc, yasm)
• Written in C++ 20

Building and running

Umbongo uses an LLVM toolchain for building, along with its own build tool for generating ninja files. freetype and nasm (yasm should also work) will also be needed. mtools and qemu are optional and are needed for building a filesystem image and running in a VM respectively.

Alternatively, umbongo can be built in a reproducible docker environment (see below).

Installing tools and dependencies

Gentoo

emerge -n \
  app-emulation/qemu \
  dev-lang/nasm \
  dev-util/ninja \
  media-libs/freetype \
  sys-devel/clang \
  sys-devel/lld \
  sys-fs/mtools

Configuring the build

To configure umbongo, call the configure.py script with a preset:

python3 configure.py \
  -B build \
  -p release \
  -e assertions \
  -e kernel_qemu_debug

Available options

Option	Description
assertions	Enables assertions project-wide
assertions_pedantic	Enables more aggressive assertions
kernel_qemu_debug	Enables kernel debug logging to port E9
kernel_stack_protector	Enables stack smashing protection for the kernel
kernel_ubsan	Enables UBSan for the kernel

Running with qemu

ninja -C build run

Installing ports

To install ports, first make sure the sysroot is up-to-date by running

ninja -C build sysroot

(note that the image and run targets automatically do this)

Then, run the port's install.sh script, passing in the build directory. For example, to install tcc, you would do

./ports/tcc/install.sh build

Building in docker

docker run --rm -u $(id -u) -v $(pwd):/src ghcr.io/yeetari/umbongo:master python3 /src/configure.py \
  -B/src/build \
  -p release \
  -e assertions
docker run --rm -u $(id -u) -v $(pwd):/src ghcr.io/yeetari/umbongo:master ninja -C /src/build