Open Source Summit Europe 2025

Linux’s current writeback infrastructure, while robust, was designed before large folios, CXL-tiered memory, and AI workloads demanding low-latency, high-throughput I/O. Today, workloads like RAG pipelines using vector databases with buffered I/O, and memory tiering on CXL, are exposing scalability limits in how the kernel handles writeback.

This talk presents a forward-looking view on evolving Linux’s writeback model. We’ll explore how the single-threaded design stalls page migration and reduces memory compaction effectiveness—affecting hugepage allocations and folio movement across memory tiers, contributing to fragmentation. On the storage side, parallelizing writeback improves throughput and responsiveness under dirty-page pressure, especially for sustained-write workloads with large memory footprints on High capacity SSDs.

We’ll also touch on early experiments within the kernel community, including efforts to make writeback more filesystem-geometry aware and parallelize it based on overwrites/new allocations.

This session invites open source community to reimagine writeback as a scalable, performance-critical component in Linux.

In the area of Linux-based real-time applications, PREEMPT_RT has become a widely embraced solution. However, it's not the only path to achieving real-time characteristics. Xenomai, utilizing a dual-kernel approach, presents a compelling alternative that merits consideration.

In this session, Richard will delve into the Xenomai project, highlighting its unique architecture and the scenarios where it excels over PREEMPT_RT and where not. Participants will gain a comprehensive understanding of how to deploy Xenomai in real-world applications through step-by-step examples.

This talk is designed not only to introduce attendees to the fundamental concepts of Xenomai but also to equip them with practical skills for implementing their own real-time solutions using this framework. Whether you’re a developer seeking to expand your toolkit or a system architect exploring different real-time options, this presentation will provide valuable insights into a versatile and powerful alternative to the conventional PREEMPT_RT route.

eBPF enables efficient tracing and monitoring of modern Linux systems. However, tools in the BCC (BPF Compiler Collection) are primarily designed for standalone use, making it challenging to adopt them directly in real-time, streaming-based observability systems.

This talk introduces a practical approach to extending BCC tools for use as data sources in system monitoring pipelines. I demonstrate an architecture that transforms BCC output into time-series data by integrating with InfluxDB, and visualizes the data using Grafana. This enables real-time tracking of kernel and user-space events such as memory allocation over time.

I also explore enhancements to existing BCC tools, such as adding options to output data in time-series–friendly formats (e.g., InfluxDB’s line protocol), enabling easier ingestion by monitoring agents. These modifications help bridge the gap between raw eBPF observability and modern telemetry systems, without compromising BCC’s standalone usability. GitHub PR (https://github.com/iovisor/bcc/pull/5281) demonstrate these improvements.

Attendees will learn to leverage BCC tools for real-time insights and contribute enhancements for broader monitoring use cases.

This talk chronicles my journey of troubleshooting a Linux kernel issue on a budget Intel GeminiLake-based Chinese mini-laptop. What began as a simple hardware purchase escalated into a two-month deep dive into the i915 GPU driver when the display mysteriously went blank during initialization.

I'll walk through the systematic troubleshooting approach: isolating the issue to the i915 driver, identifying the kernel configuration options triggering the problem, and developing a practical patch that bypasses problematic GPIO pin activation sequences. Along the way, I'll share surprising discoveries about hardware compatibility, kernel development complexity, and the limitations of AI tools when facing real-world Linux challenges.

This presentation is designed for Linux enthusiasts and IT professionals curious about kernel troubleshooting. Attendees will leave with practical knowledge about GPU driver internals, confidence that such issues are solvable without specialized expertise, and inspiration to tackle their own hardware compatibility challenges.

The DevRel Foundation is a vendor-neutral initiative aimed at building shared understanding, resources, and best practices to elevate the professional practices of developer relations across industries. In response to recent challenges, including industry layoffs and the evolving landscape of DevRel roles, the foundation has established working groups to tackle pressing issues such as metrics and reporting, career support, and resource aggregation. This session offers attendees the opportunity to connect with thought leaders in the DevRel Foundation community, engage in discussions about current challenges, and explore ways to get involved in shaping the future of the developer relation foundation. Key questions to explore include: What strategies can showcase the return on investment of DevRel initiatives to stakeholders? What skills are becoming important in the evolving DevRel landscape, and how can professionals acquire them? How can engaging with the DevRel Foundation’s working groups enhance your career?

The final output will result in a report capturing the roundtable's insights and recommendations authored by the attendees, hosted under Developer Relations' Foundation GH organization under CC-BY 4.0.