Tracing track

Friday, October 17, 2014 from 9:00amNoon
Room 2

The tracing microconference will not be about presentations. It will be a plenary meeting of all participants in order to try and solve the issues within the tracing community.

Currently, there are many players: ftrace (with trace-cmd and kernelshark), perf (both kernel and userspace), systemtap, LTTng. There are also a need to get a programmable dynamic filtering into the kernel. This will require ktap and the BPF system coming together.

Microconference Leader

Steven Rostedt

Sessions for this track

* Correlating timestamps in user and kernel space

We need a way of timestamping perf(ormance) data generated in userspace.
Paweł Moll

* Linux Tracing Strategy

The current state of tracing, integration between and of the tracers, and where we should be going.
Brendan Gregg

* Rich probe filtering and reporting with variable locations and types

How to combine various mechanisms (CTF, SDT, DWARF, etc.) used by various tools to provide rich filtering and reporting of events.
Mark Wielaard

* Sharing kernel tools code

Making perf tools generic code available as a tools library.
Jiri Olsa

* Use of Common Trace Format (CTF) among different tracers

Next steps to share common tracing tools via common trace format
Mathieu Desnoyers

Proposals for this track

* dmesg buffer: limitations of the original kernel tracing tool

Discuss limitations of dmesg buffers to triag ChromeOS kernel crashes.
Tracing 08/26/2014
Grant Grundler

* Processor (hardware) trace

We have ARM CoreSight and Intel PT patches on the mailing lists. We need to find common ground for them.
Tracing 10/01/2014
Paweł Moll

* Tracing on large scale infrastructure

How to best use tracing tools on severals thousands of servers
Tracing 10/06/2014
Yannick Brosseau