Adding Inter-event Capabilities to Linux Kernel Trace Events
Talkat Embedded Linux Conference + OpenIoT Summit North America 2017 · on Wednesday, 2017-02-22· at 16:20· Grand Ballroom I/II·
summary
Patch set adds inter-event capabilities to the Linux trace event subsystem so latencies are computed inside the kernel without external scripts or runtimes. It implements synthetic events as containers, reuses hist triggers as hash tables for variables, and exports common_timestamp for every event. Actions trigger synthetic events on matches. Tested with cyclictest to produce wake‑up latency histograms (observed 1–26 µs with a peak at 2–3 µs); patches are posted to LKML and receive review and iterated fixes.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:58Linux Trace Events
- 3:20Enabling Events
- 4:53Triggering Events
- 9:30Latency
- 14:25The Problem
- 17:55The Variables
- 22:07Synthetic Events
- 25:19Creating a Synthetic Event
- 26:16Sending a Synthetic Event
- 29:36Recap
- 31:10Conclusion
- 32:49Future Plans
- 35:26Wrap Up
- 37:04Recursive Events
- 38:14Arithmetic
related talks
41mSynthetic Events and Tracing Latency Within the Kernel
46mMeasuring and Summarizing Latencies using the Trace Event Subsystem
48mCyclictest latency, find out why with tracing
55mAdvanced Tracing Features to See What Your Application is Waiting On
38mLTTng: Beyond Ring-Buffer Based Tracing
35mLinux Tracing updates
49mDebugging Methodologies for Realtime Issues in Linux Systems
52mTracking Down What Happened: A Comparison of Tracing Solutions