“Is Upstream Really Enough?” — Practical Realities of Using eBPF in Long-Term Supporte...
summary
Upstream eBPF innovation is not enough for LTS and embedded products: kernel internals and tracepoint behavior change, verifier incompatibilities and shifting kfuncs force expensive backports, and libbpf / libclang hell breaks embedded CI and toolchains while some vendors outright ban bpf() for security. Proposal: publish an eBPF Embedded Profile — a narrow, tested subset of APIs, stable kfunc list, per-use-case kernel configs and deployment guidance to reduce backport cost, improve runtime predictability, and enable certification for production use.
AI-written summary · about
chapters (7)
- 0:00Introduction and background
- 2:33Goal of the talk
- 4:20Embedded system requirements
- 5:51Challenges of eBPF in embedded
- 11:42Proposed eBPF embedded profile
- 13:53Next steps and collaboration
- 15:38Key takeaways and conclusion
related talks
32mChallenges of Deploying eBPF-based Tracing in Embedded Systems, and Alternatives in Embe...
20mClosing Perspectives
29meBPF on the Rise - Getting Started
32meBPF: The Next Power Tool of SREs
33meBPF: exploring use case of BPF kernel infrastructure
40mPanel Discussion: eBPF: A New Era in Cloud Infrastructure Tools
14mWill eBPF Save Us From the Next Global Outage?
29mBuzzing Across the Cloud Native Landscape with eBPF