home.social

#preempt_rt — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #preempt_rt, aggregated by home.social.

fetched live
  1. A year ago I was experimenting with Debian netinst and live-build, trying to understand boot, init, and live systems.

    Today BashCore is a set of minimalist Debian-based live ISOs (CLI/GUI, optional PREEMPT_RT).

    In 2026 the focus shifts to Buildroot: reproducible systems, RT Linux, and minimal OS design for real workloads.

    Happy New Year to everyone!

    bashcore.org

    #Linux #Buildroot #Debian #PREEMPT_RT #RealTimeLinux #Embedded

  2. 🚀 Rock-solid Real-Time performance on Raspberry Pi 4!

    Just stress-tested via SSH a custom #Buildroot (Kernel 6.12.61-v8 #PREEMPT_RT). Even with a massive memory hog running (stress-ng --vm 2 --vm-bytes 1G), the results are insane:

    🔹 Avg Latency: 14µs
    🔹 Max Latency: 42µs
    🔹 Jitter: Virtually zero

    This #RPi4 is now a deterministic beast. 🐧🛠️ 64-bit RT Linux at its best!

    #Linux #RealTime #EmbeddedLinux #RaspberryPi #Kernel #Performance #SystemStability

  3. Built a custom Raspberry Pi 4 RT OS using Buildroot.

    Manually downloaded the Raspberry kernel (6.12.61) and PREEMPT_RT patch (6.12.57), applied locally and produced a bootable SD.

    It works.

    No GUI, CLI tools only, deterministic boot and real-time behavior.

    Learning by doing 🤓

    #Linux #Buildroot #RaspberryPi #PREEMPT_RT #RealTime #Embedded #NoGUI #CLI #Nerd

  4. Запуск приложений в Linux c использованием политик планирования «реального» времени

    В октябре 2025 года в Нижнем Новгороде на форуме "Промышленная автоматизация: переход на открытую АСУ ТП" с коллегами зашел интересный разговор о Linux с патчем реального времени. В разговоре затронули разные интересные моменты: работа через гипервизоры, выделения ядер и отличия от QNX. Договорились о том, что я напишу небольшой гайд, рассказывающий что и как можно сделать, чтобы оптимизировать время отклика. Отдельно отмечу, что задачу доказать, что Linux с патчем PREEMPT_RT является операционной системой жёсткого реального времени перед собой не ставлю. На мой взгляд, она такой не является, так сказать by designe, в отличие от тех же QNX , VxWorks и ThreadX.

    habr.com/ru/articles/966210/

    #linux #preempt_rt #репка_пай #repkapi

  5. Halfway point (24h) in the Atom N450 Real-Time endurance test! The tiny netbook is stable and deterministic under 100% CPU load.

    Key Results:

    ✅ System stability confirmed (0.0% CPU usage for kworkers/rcu).

    ⏱️ Max Jitter remains extremely low after 20 hours: 174μs (microseconds).

    This proves the PREEMPT_RT kernel is effectively managing the worst-case scenario!

    #RealTimeLinux #PREEMPT_RT #AtomN450 #BashCoreZ

  6. Endurance test on a 2009 Atom N450 netbook! 🛠️
    ​This Acer Aspire One always failed on standard Linux due to a frustrating #kworker CPU spike, making it unusable. We're challenging that with #BashCoreZ_RT for 48 hours.
    ​Goal: Prove our extreme #PREEMPT_RT optimization can permanently eliminate this hardware bug and revive the PC. Superior software rescues vintage hardware. Wish us stability! 🤞
    #LinuxRT #KworkerFix #VintageTech #LowLatency

  7. The **BashCoreZ RT** supremacy continues! We just tamed a 2009 Intel Atom N450 netbook (2GB RAM).

    MAX Jitter hit an impressive **112µs** (-t1). The difference when scaling to 16 threads (-t16) was only 42µs!

    This confirms it: our zero-bloat #PREEMPT_RT architecture minimizes scheduling overhead on even the weakest legacy CPUs.

    We're bringing guaranteed #RealTime performance to obsolete hardware. BashCoreZ RT makes ancient PCs viable again. Zero excuses. 👑

    #LinuxRT #LowLatency #Netbook

  8. Checkmate. The final #BashCoreZ #PREEMPT_RT test is in:

    On a budget 2-core Celeron N4500 laptop (250€), Max Jitter hit an insane **39µs** after isolating network noise!

    This proves it: Hardware specs count for little. The clean, zero-bloat kernel is EVERYTHING. We're delivering near Hard Real-Time performance with minimal cost.

    29µs on i7 vs. 39µs on Celeron? The difference is negligible. BashCoreZ RT is the definitive, free Real-Time solution. 👑

    #LinuxRT #LowLatency #Celeron #MITLicense

  9. We didn't just win; we dominated. The ultimate #BashCoreZ #PREEMPT_RT test results are here.

    On a clean, budget MSI i7 laptop (500€!), Max Jitter hit an insane **29µs**! That's the true potential of our "zero bloat" kernel 6.12 base.

    (Context: Our previous "worst-case scenario" test, with some apps running, capped at only 239µs. Competitors struggle to hit that clean!)

    BashCoreZ delivers Hard Real-Time performance without the huge price tag. Get building. 👊🏻

    #LinuxRT #Embedded #Trixie

  10. BashCoreZ just proved it's the ultimate Real-Time base! 🚀

    Tested the ultra-minimal #BashCoreZ (Debian 13 + Kernel 6.12 #PREEMPT_RT) using #cyclictest, remotely piloted from a #BashLabOS host.

    Result: Max Jitter stayed under 240µs! That’s incredible determinism for embedded, robotics, and industrial control.

    The "zero bloat" foundation truly eliminates noise.

    bashcore.org

    #LinuxRT #EmbeddedLinux #Trixie #Debian

  11. hdblog.it/linux/articoli/n5999
    Una delle novità principali del Kernel Linux 6.12 è l'integrazione del supporto #PREEMPT_RT una funzionalità che migliora notevolmente le prestazioni ...
    #UnoLinux

  12. 2/ Side note for users of :

    6.12 is now available for all Fedora releases in the "stable" of my vanilla repositories.

    For install instructions, see copr.fedorainfracloud.org/copr and fedoraproject.org/wiki/Kernel_

    In addition to the regular kernel package that copr as of today started shipping a dedicated kernel which has enabled. Install it by running:

    $ sudo dnf install kernel-rt

  13. 2/ #realtime / -rt / #PREEMPT_RT #kernel packages for #Linux-next are now available as well: copr.fedorainfracloud.org/copr

    Side note: instead of 2 to 3 hours the daily builds now take 3 to 6. 😕

    But whatever. 😀

  14. 🥳 My #kernel vanilla repositories for #Fedora #Linux[1] started shipping[2] kernels with #realtime aka #PREEMPT_RT support. 🥳

    [1] fedoraproject.org/wiki/Kernel_
    [2] 6.12-pre and later only obviously, as the latest stable series does not support it out-of-the-box

  15. With the new kernel hitting mainline I wrote an article on the importance of kernel parameter tuning to reduce latency and jitter for

    dantalion.nl/2024/09/29/linuxc

  16. v6.12-rc1-rt1 is out:

    lore.kernel.org/all/2024093015

    If you wonder what's still in the RT-Tree now that support is upstream, checkout git.kernel.org/pub/scm/linux/k

    In short: about 40 patches that among others switch 8250 to nbcon console, add support for ARM and PowerPC, or fix some drivers (like i915).

  17. 6.12-rc1 is out:

    lore.kernel.org/lkml/CAHk-=wiw

    '"Despite conference travel (both for me and several maintainers), things seemed to go mostly fairly normally. There's a couple of notable new features in here: For one thing, is now mainlined and enabled as a config option (you do need to enable "EXPERT" to get the question). For another, also got merged.

    […]

    Let's get the testing and calming down period started, ok?

    Linus"'

  18. Linux 的 RT (Real-Time) 版本 PREEMPT_RT

    在 Hacker News Daily 上看到的,Linux kernel 把 PREEMPT_RT 這個 RT patch 併入 official repository 了:「Real-time Linux is officially part of the kernel after decades of debate」。

    這類需求可以在 RTOS 裡面看到完整的說明,大家常用的特性是確保反應時間這點,像是對汽車自動

    blog.gslin.org/archives/2024/0

    #Computer #Linux #Murmuring #OS #Software #atomic #kernel #linux #output #patch #preempt_rt #printk #realtime #realtime

  19. 2/ It's done! 🥳

    A few minutes ago (aka: support) support landed in mainline, as Linus honored the PR that Tglx handed him yesterday in person (see first toot in thread): git.kernel.org/torvalds/c/baeb

    Congrats and thx to everyone who was involved in this 20 year long journey!

  20. The 6.12 pull request for is handed to Linus. Likely the first PR ever submitted in printed form.

  21. aka the support for hit -next today! 🥳 👏

    git.kernel.org/pub/scm/linux/k

    Still remains to be seen if those changes really makes it upstream during the 6.12 merge window and stick:

    * It afaics is still unclear if some prerequisites will be submitted for 6.12: lore.kernel.org/all/87ed5za2oj

    * Troublemaking changes sometimes are reverted before the final, which in fact happened to some of printk prerequisites a while ago: lore.kernel.org/lkml/202206231

  22. patches to allow enabling (aka proper support) for x86, ARM64, and Risc-V were posted for review:

    lore.kernel.org/lkml/202409061

    "The printk bits required for PREEMPT_RT are sitting in linux-next. This was the last known roadblock for PREEMPT_RT. The RT queue has additionally the "atomic console" for the 8250 UART which is not yet in linux-next.[…]"

    See also these earlier other related toots:

    fosstodon.org/@kernellogger/11
    fosstodon.org/@kernellogger/11

  23. v6 of the patch-set "add threaded printing + the rest" which "provides the remaining pieces of the rework."[1] is now in -next – but as of now it's not yet clear if this will be merged for 6.12.

    lore.kernel.org/all/ZthvGoJE26

    [1] e.g. the last big missing piece for support through in the mainline . But note, this series does *not* provide an nbcon console driver. That will come in a follow-up series.

  24. John posted v4 of the patch-set "threaded console printing as well as some other minor pieces" which "provides the remaining pieces of the rework." -- e.g. the last big missing piece for support through in the mainline . But note, "this series does *not* provide an nbcon console driver. That will come in a follow-up series."

    lore.kernel.org/all/2024082704

    This comes hot on the heels of the "write_atomic() printing" entering -next: fosstodon.org/@kernellogger/11

  25. The improved series "wire up write_atomic() printing" is now in -next again and thus slated for inclusion in 6.12. This is one of the last important bits for proper support with in mainline – but not the last, as this series does not include threaded printing or nbcon drivers, as those will come later: lore.kernel.org/all/ZsWxpVG8uZ

    A earlier version of the patch set was supposed to go into 6.11, but Linus was unhappy: lore.kernel.org/lkml/CAHk-=whU

  26. Linus did not pull the latest printk changes for 6.11, as he did not like the what some of the changes required for proper aka support do:

    lore.kernel.org/all/CAHk-%3Dwh

    "'
    > The messages are flushed at the end of the emergency section to allow storing the full log (backtrace) first.

    What? No.

    One of the historically problematic situations is when a recursive oops or a deadlock occurs *during* the first oops.

    […]
    '"