Loading…

Sign up or log in to bookmark your favorites and sync them to your phone or calendar.

Linux Systems [clear filter]
Tuesday, October 27
 

13:00 GMT

DTrace: Leveraging the Power of BPF - Kris Van Hees, Oracle Corp.
BPF and the overall tracing infrastructure in the kernel has improved tremendously and provides a powerful framework for tracing tools. DTrace is a well known and versatile tracing tool that is being re-implemented to make use of BPF and kernel tracing facilities. The goal of this open source project (hosted on github) is to provide a full-featured implementation of DTrace, leveraging the power of BPF to provide well known functionality

The presentation will provide an update on the progress of the re-implementation project of DTrace. Kris will share some of the lessons learnt along the way, highlighting how BPF provides the building blocks to implement a complex tracing tool. He will provide examples of creative techniques that showcase the power of BPF as an execution engine.

Like any project, the re-implementation of DTrace has not been without some pitfalls, and Kris will highlight some of the limitations and unsolved problems the development team has encountered.

Speakers
KV

Kris Van Hees

Consulting Software Engineer, Oracle Corp.
Kris Van Hees works for Oracle Corp. He works primarily on tracing and debugging tools. Previously Linux projects (incl. with former employers) include OpenAFS, zLinux, and DTrace. His current project is the re-implementation of DTrace on top of Linux kernel tracing features like... Read More →



Tuesday October 27, 2020 13:00 - 13:50 GMT
Linux Systems Theater
  Linux Systems, Tracing
 
Wednesday, October 28
 

17:15 GMT

Customized Trace Data Visualization with KernelShark - How to Write User Plugins. - Yordan Karadzhov, VMware
KernelShark is a front-end reader of tracing data and its data visualization capabilities have been proven very useful. Nevertheless, there are cases when the user has specific needs, that go beyond what is provided by the built-in visualization model. Although, the user customization was one of the key features incorporated in the design of KernelShark v1, the possibility to write plugins was not taken advantage by the KernelShark users. We believe that this will change with the release of KernelShark 2.0. The new version includes a substantial improvement to the infrastructure for plugins. The plugin development process was greatly simplified, while at the same time the user has more ways to customize. This talk will be a brief overview of how to create your own plugins and will demo some of the new key features in KernelShark 2.0, that are implemented in the form of plugins.

Speakers
avatar for Yordan Karadzhov

Yordan Karadzhov

Open source engineer, VMware
Yordan Karadzhov has more than 12 years of experience as experimental physicist, includes a Ph.D. in particle physics. During this period Yordan worked in some of the world's largest physics laboratories, like CERN, FermiLab and RAL, developing software for particle physics experiments... Read More →



Wednesday October 28, 2020 17:15 - 18:05 GMT
Linux Systems Theater
  Linux Systems, Tracing

17:30 GMT

Lightning Talk: Unravelling the Linux Kernel Using BPF Raw Tracepoints - Abhijit Singh, Uber
This talk aims at presenting the potential of BPF raw tracepoints. In a static tracepoints present in the Linux kernel, a dynamic tracer like ftrace or kprobe+bpf, might not get desired information about arguments passed in the tracepoint(e.g. many fields related to the task_switch struct aren't accessible such as nvcsw and nivcsw in the sched_switch tracepoint). With the introduction of BPF raw tracepoints, we can create a BPF program which can investigate each argument being passed to the tracepoint upon its invocation. The talk will commence with the rationale to introduce a raw BPF tracepoint. It would then present the signature of the BPF_RAW_TRACEPOINT api. It would also present a couple of example BPF programs which use raw tracepoints. 1) a program which prints the arguments being passed to the openat system call 2) a program which calculates total involuntary context switches, voluntary context switches over time for a system The talk would also highlight the low performance overhead of raw tracepoint as compared to other tracing infrastructure like kprobe+bpf.

Speakers
avatar for Abhijit Singh

Abhijit Singh

Software Engineer, Uber
I'm a software engineer, currently working in Uber. Previously, I used to work as a Performance Engineer at Azul Systems where I conducted thorough performance analysis of JVM applications and the underlying systems. I'm a keen systems performance enthusiast and follow Linux kernel... Read More →


Wednesday October 28, 2020 17:30 - 17:40 GMT
Linux Systems Theater
  Linux Systems, Tracing
 

Twitter Feed