virtio-accel Linux kernel module

The virtio-accel Linux kernel module implements VirtIO-based transport for acceleration operations. It is meant as a lightweight alternative to vSock for vAccel VM execution.

Following the split-driver VirtIO model, the transport functionality is implemented in two parts: a) this kernel module, and b) a QEMU driver

Prebuilt artifacts

You can get the latest VM artifacts, including virtio-accel and a Linux kernel image, using:

wget https://s3.nbfc.io/nbfc-assets/github/vaccel/virtio-accel/rev/main/${ARCH}/release/virtio-accel-latest-bin.tar.xz
wget https://s3.nbfc.io/nbfc-assets/github/vaccel/virtio-accel/rev/main/${ARCH}/release/virtio-accel-latest-linux-image.tar.xz

by setting ARCH=x86_64 or ARCH=aarch64 depending on your architecture.

There is also a Docker image with QEMU pre-installed:

docker pull harbor.nbfc.io/nubificus/qemu-vaccel:${ARCH}

Build virtio-accel

To build the virtio-accel module you will need the Linux kernel headers for your VM's kernel. To install the headers on a VM with a Debian-based distro you can use:

sudo apt install linux-headers-$(uname -r)

Get the plugin:

git clone https://github.com/nubificus/virtio-accel
cd virtio-accel

To then build virtio-accel with meson:

meson setup build
meson compile -C build

You can also configure a custom kernel source directory using the kdir option, ie:

meson setup --reconfigure -Dkdir=/path/to/kernel/source build

A plain make implementation is also available:

make modules

or

make modules KDIR=/path/to/kernel/source

Alternatively, artifacts to use for a VM can be built with:

ninja gen-vm-artifacts -C build

using an already configured meson build directory. This target requires a working Docker installation.

Build QEMU

To use the virtio-accel module you will also need a compatible QEMU build.

To fetch and build QEMU:

git clone https://github.com/cloudkernels/qemu-vaccel
mkdir qemu-vaccel/build
cd qemu-vaccel/build
../configure --target-list=${ARCH}-softmmu --enable-virtfs && \
make -j$(nproc)