FOSDEM 2026 Recap: eBPF Observability on RISC-V

At FOSDEM 2026, DeepComputing Founder and CEO Yuning Liang joined Bruce Gain to present a session in the eBPF track exploring the current state of eBPF observability on RISC-V. The talk focused on a practical question that many developers are starting to ask: How ready is the RISC-V ecosystem for modern observability tools? As RISC-V systems expand into servers, edge computing, and AI platforms, observability becomes essential for debugging, performance tuning, and production reliability. eBPF has become the industry standard for Linux observability — but enabling it on a rapidly evolving architecture like RISC-V comes with unique challenges.

Why eBPF Matters for RISC-V

eBPF is widely used across cloud infrastructure, networking, security, and performance monitoring. For RISC-V to become a competitive platform in these domains, full eBPF functionality is critical.

The session highlighted how eBPF enables:

• Low-overhead system tracing
• Performance profiling and debugging
• Networking and security observability
• Production monitoring in modern infrastructure

Without strong observability tooling, deploying RISC-V systems at scale becomes significantly harder.

What Works Today

One of the encouraging messages from the session was that significant progress has already been made.

Developers can today:

• Run many eBPF programs on RISC-V Linux kernels
• Use common tooling such as bpftool and tracing frameworks
• Perform basic observability tasks on modern RISC-V platforms

This progress reflects strong upstream work across the Linux kernel and open-source communities.

What Still Breaks

However, the session also provided an honest look at the gaps that remain.

Some of the challenges discussed included:

• Kernel feature parity compared with mature architectures
• Toolchain and verifier differences
• Missing or incomplete functionality in certain subsystems
• Performance and stability inconsistencies
• Testing coverage limitations

These issues are not unusual for a growing architecture — but identifying them clearly is essential for improvement.

Collaboration Across the Ecosystem

One of the key takeaways from the session was that enabling eBPF on RISC-V is a community effort.

Progress depends on collaboration between:

• Kernel maintainers
• Toolchain developers
• Hardware vendors
• Distribution maintainers
• End users and testers

DeepComputing continues to contribute through hardware platforms, validation work, and upstream collaboration to help close the observability gap.

Watch the Session