c0365df3ca
Build OFED drivers and userspace components against the kernel source tree similar to Intel's NIC drivers. OFED installers create Debian packages of their own tageting the kernel version defined in the build invocation if DKMS is omitted. Script builds with supporting components for VPP to permit handoff of function to the underlying hardware as appropriate. Updating the version is fairly trivial along with adding patching as needed to handle kCFI and hardening measures as they are introduced. Testing: Tested against GCC-built Linux Hardened kernel with the various additions from PR 132 - sustained line-rate testing against 4x100g links on a single machine at a hair below 200g for each LACP pair.
About
VyOS runs on a custom Linux Kernel (which is 4.19) at the time of this writing. This repository holds a Jenkins Pipeline which is used to build the Custom Kernel (x86_64/amd64 at the moment) and all required out-of tree modules.
VyOS does not utilize the build in Intel Kernel drivers for its NICs as those Kernels sometimes lack features e.g. configurable receive-side-scaling queues. On the other hand we ship additional not mainlined features as WireGuard VPN.
Kernel
The Kernel is build from the vanilla repositories hosted at https://git.kernel.org. VyOS requires two additional patches to work which are stored in the patches/kernel folder.
Config
The Kernel configuration used is x86_64_vyos_defconfig
which will be copied on demand during the Pipeline run into the arch/x86/configsi
direcotry of the Kernel source tree.
Other configurations can be added in the future easily.
Modules
VyOS utilizes several Out-of-Tree modules (e.g. WireGuard, Accel-PPP and Intel network interface card drivers). Module source code is retrieved from the upstream repository and - when needed - patched so it can be build using this pipeline.
In the past VyOS maintainers had a fork of the Linux Kernel, WireGuard and Accel-PPP. This is fine but increases maintenance effort. By utilizing vanilla repositories upgrading to new versions is very easy - only the branch/commit/tag used when cloning the repository via Jenkinsfile needs to be adjusted.