Tuning Real-Time Linux for ARM SoCs
summary
Texas Instruments debugs Real-Time Linux on heterogeneous ARM SoCs where TrustZone, GPU stacks and firmwares share I/O, DDR and CPU bandwidth and can unpredictably degrade latency. Testing combines stress-ng, perf, cyclictest and rtla with GPIO waveform checks using an oscilloscope persistence capture to measure min/max latency; tracer work finds surprises like an RNG secure-world spin bug that appears about one in a million. Practical fixes are to remove unnecessary kernel modules, disable unused subsystems, and tune tick/nohz and IRQ priorities, yielding about two thirds latency reduction in some cases.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 1:01Real-time Linux requirements
- 2:06Latency and determinism
- 3:31Debugging latency issues
- 5:16Automated stress testing
- 7:27Hardware vs software priorities
- 8:32Kernel size optimization
- 10:36Key optimizations summary
- 10:56Q&A
- 12:22Standardizing real-time tests
- 13:49Kernel growth challenges
- 14:32Reproducible testing
- 15:57Reference builds and CI
- 17:02Baseline measurements
- 18:05Real-time networking
- 19:36GPIO latency
- 20:39Minimizing latency deviation
- 21:34Historical context
- 22:18Real-time on laptops
- 22:59Stress test corner cases
- 24:04Debugging firmware/drivers
- 24:49Automated firmware testing
- 25:10CI and manual workflows
related talks
26mTuning Real-Time Linux for ARM SoCs
44mReal-Time Linux on Embedded Multicore Processors
40mTips, Tricks, and Gotchas for Linux Real-Time Tuning
49mDebugging Methodologies for Realtime Issues in Linux Systems
47mReal-Time Linux: What Is Next?
54mIntroduction to Realtime Linux
50mrtla timerlat: Debugging Real-Time Linux Scheduling Latency
27mAccelerated Mainline Linux Development Ahead of SoC Availability