Open Source

Meinrad Happacher,

End of the Linux Real-Time Dilemma

For 20 years, there have been software extensions that grant real-time capabilities to the Linux kernel. Carsten Emde, General Manager of OSADL, now reports a breakthrough on this front.

Carsten Emde: »The Linux kernel will now require far less post-release adaptation for real-time, a significant relief for everyone involved.« © OSADL

Mr. Emde, after all this time, is there still anything groundbreaking left to achieve in terms of real-time Linux?

Carsten Emde: It’s true that there’s been a software extension for 20 years that altered many aspects of the Linux kernel to give it real-time capabilities. That was a technical revolution since many experts thought it impossible. But the question remained whether it would be feasible to align these numerous modifications with the quality and maintainability standards of the original Linux kernel, so that they could gradually be incorporated. And as of September 20, 2024, shortly after 6 a.m., we can now answer this second question with a resounding »Yes.« At that moment, the last remaining changes of the real-time extension were released for the original Linux kernel.

With the December release, it will, for the first time, be possible to configure the original Linux kernel directly for real-time capabilities - and that’s the long-awaited milestone of epic significance: a spectacular open-source development project spanning over 20 years has reached a critical goal!

Are the real-time capabilities of the Linux kernel now improved?

Advertisement

No. By and large, a Linux kernel equipped with the real-time extensions 20 years ago offered similar real-time properties as today’s real-time-configured original Linux kernel. The key difference lies in the maintenance effort required for the extensions. Previously, this off-tree development had a significant drawback: it wasn’t synchronized with the efforts of other kernel developers. For each new kernel version—four times per year over 20 years, totaling 80 updates—the extensions had to be adapted to the new kernel. Sometimes, this was simple when little or no shared code was affected. However, it could also be challenging, or even impossible, if the new concepts introduced were incompatible with real-time requirements. When adjustments proved impossible, the challenge was to collaborate with other developers to find a mutually acceptable solution. The elimination of this often frustrating and redundant adaptation process is a great relief for everyone involved.

If such a large part of the work is no longer needed, will the Linux real-time developers now need to find new projects?

No, definitely not. The developers can now fully focus on the current and future challenges associated with real-time requirements. There is already a list of special use cases where further improvements in real-time behavior and performance of real-time systems would be beneficial. And just as new technologies have continuously emerged in the past, requiring adjustments to the real-time extensions, the same will be true in the future. The existing approach will continue, whereby proposals for extensions are maintained separately and, once they reach sufficient maturity, are incorporated into the original Linux kernel. One thing is certain: real-time developers will not be bored anytime soon.

OSADL at SPS 2024: Hall 6, Stand 454

  • Xing Icon
  • LinkedIn Icon
Advertisement
Advertisement

You might also be interested in

Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Subscribe to our newsletter
Advertisement
Back to home