Open Source Summit Europe 2025

Embedded systems vary wildly in purpose, characteristics and usage. They are therefore faced with very different adversaries, that engineers and designers must take into account to achieve a secure system or product.
The tradeoff between security and convenience can only be properly arbitrated by evaluating the benefits of each security measure against the cost of its implementation, and the added complexity to the system, which will eventually be surfaced to the user.

In this talk, we will explore the security guarantees of various
security measures popular in embedded Linux systems: Secure Boot, Disk encryption, dm-verity, Secure Enclaves as well as someapplication-level security features from the point of view of what they bring to a system, and what they cost to the implementer or theuser. Ultimately, we will present what type of systems they are a good fit for, and which systems are probably better off not using them.

This session dives into the hardware and software design decisions behind restructuring a PMIC driver to support multiple device families with minimal upstream churn. We'll explore techniques for scalable driver design, lessons learned from managing variations across similar devices, and strategies for maximizing code reuse across subsystems. Real-world examples of accepted and discarded approaches will be shared. The goal is to spark a community discussion around best practices for future multi-device driver development across various hardware families.

The ODM ecosystem — such as the one found in Asia — offers a rich selection of pre-designed hardware intended for white-labelling. This hardware can be produced and shipped with short lead times. However, a typical weakness of these devices is their software, which usually consists of outdated kernels and poorly documented components. Wholesale replacement of the system’s software stack while leveraging the hardware offers a path toward creating robust devices that can sidestep many of the complexities of manufacturing a fully custom design.

In this session we share tips and learnings from our experience creating a replacement software stack using Yocto for an existing white-label IoT gateway device. With this approach, we were able to take a new device with an up-to-date Linux stack to production in under 6 months. 

The session will cover topics including: choosing of an appropriate hardware platform, coordinating with the hardware vendor, setting up an efficient development environment using similar hardware and simulators, and using the system’s OTA update infrastructure to create a factory image that’s compatible with an ODM’s existing manufacturing and testing flows.

As launch cadence increases and development cycles tighten, the space industry turns to open source to meet the moment. Enter Space Grade Linux (SGL) — an initiative under the ELISA Project aimed at creating a reusable, safety-aware Linux foundation for spaceflight systems.

This talk will introduce the goals and current status of SGL, highlighting three foundational focus areas:
1. Kernel Configuration – Defining a shared starting point for space-focused Linux systems, emphasizing predictability, determinism, and traceability.
2. Booting into Linux: Exploring the safety-critical implications of system bring-up and strategies for improving reliability in space-grade deployments.
3. Userspace Strategy – Discussing early-stage decisions around minimal runtime environments, supervision, and what a safe, maintainable userspace might look like.

Attendees will get a hands-on overview of what’s already available in the GitHub repository, including a Yocto-based reference implementation and working kernel configuration. More importantly, they’ll learn how to get involved — through technical contributions, architecture discussions, or community collaboration.

WiFi is a cornerstone of modern connectivity, and Linux powers countless devices that rely on seamless wireless communication. But how does a WiFi driver actually work within the Linux kernel? In this talk, we will explore the process of supporting a Wifi device in Linux, focusing on its integration with the mac80211 and cfg80211 subsystems. We’ll cover key aspects such as hardware initialization, packet transmission, regulatory compliance, power management, and more. Additionally, we’ll discuss how userspace tools like iw, hostapd, and wpa_supplicant interact with kernel drivers via nl80211. By the end of this session, attendees will have a better understanding of the Linux WiFi stack and the skills needed to develop, debug, and extend WiFi drivers effectively, whether they want to implement a new one or improve an existing one.

Modern cars use a high integration architecture with a smaller amount of powerful Electronic Control Units (ECUs). They easily accommodate more than 150 different services and are developed by more than 1000 developers in dozens of companies across the world. Ensuring security in such an environment is a challenge in multiple dimensions: Attack surface, overall system complexity and communication.

In this talk, I want to share our insights from building three generations of Yocto-based infotainment systems. How do we ensure basics such as the least privilege principle and privilege separation? What is our approach to trusted computing with secure boot and dm-verity? What are our lessons learned to enable SELinux on this large scale? Which best-practices do we have so that the system can be developed, tested and debugged in a (security) configuration that is close to the final product?

Furthermore, I would like to give positive examples how a good open-source ecosystem can support our cause of building secure embedded Linux systems as well as some impulses which improvements could help us a lot.

Power management in heterogeneous SoCs with multiple asymmetric multi processing cores running different operating systems needs a coordinated approach to attain SoC level low power states.

In order to choose a suspend to RAM mode, requirements from applications running on co-processors need to be communicated and honored.

This talk shall cover how the Linux remoteproc driver should implement the system suspend resume functions that will coordinate with firmware running on a remote core during suspend and resume of the system.

This talk will give attendees the understanding of what it takes for a heterogeneous SoC to leverage the remoteproc subsystem to implement graceful suspend and resume of remote cores.

We shall also look at the existing state of remoteproc driver in kernel, areas of improvements and optimization, and blockers we are facing in upstreaming suspend resume functionality in the TI’s remoteproc driver.