Hi Retina team,
I maintain bpfcompat (Apache-2.0): it boots real distro kernels in disposable QEMU/KVM VMs, load/attach-validates a compiled .bpf.o, and reports a per-kernel pass/fail matrix with classified failure reasons (missing BTF, CO-RE relocation, unsupported map/program/attach type).
Retina's eBPF plugins (conntrack, dropreason, dns, …) run across AKS and a wide range of other distros/kernels — exactly the compatibility surface bpfcompat targets. I'd be happy to run Retina's compiled objects across a kernel spread and share the matrix as complementary evidence / a possible non-blocking CI lane.
Context: a Falco modern_bpf reference matrix and an enterprise/backported-kernel matrix (RHEL 8/9/10, Oracle UEK, Amazon Linux 2 incl. the no-BTF 4.14, openSUSE — 14/14). Live demo.
Would a published per-kernel matrix for the plugins (or a CI lane that gates on it) be useful? Happy to put one together.
Independent test of a public project; not affiliated with or endorsed by Microsoft/Retina.
Hi Retina team,
I maintain bpfcompat (Apache-2.0): it boots real distro kernels in disposable QEMU/KVM VMs, load/attach-validates a compiled
.bpf.o, and reports a per-kernel pass/fail matrix with classified failure reasons (missing BTF, CO-RE relocation, unsupported map/program/attach type).Retina's eBPF plugins (conntrack, dropreason, dns, …) run across AKS and a wide range of other distros/kernels — exactly the compatibility surface bpfcompat targets. I'd be happy to run Retina's compiled objects across a kernel spread and share the matrix as complementary evidence / a possible non-blocking CI lane.
Context: a Falco modern_bpf reference matrix and an enterprise/backported-kernel matrix (RHEL 8/9/10, Oracle UEK, Amazon Linux 2 incl. the no-BTF 4.14, openSUSE — 14/14). Live demo.
Would a published per-kernel matrix for the plugins (or a CI lane that gates on it) be useful? Happy to put one together.
Independent test of a public project; not affiliated with or endorsed by Microsoft/Retina.