Getting Started With BPF
Details
BPF is the technology in the Linux kernel, evolved from BPF, then eBPF, that has a wealth of capabilities for observing behavior in the Linux kernel and in user processes. Data structures, function calls/returns, histograms collected in the kernel, and more.
We’ll describe the breadth of observations possible – CPU, network, memory, file systems, disk devices, ...
We will also discuss the BPF mechanism and then examine available tools that use BPF and demonstrate how you can write your own simple tools.
Making use of BPF tools is essential these days for anyone needing to get insights into performance or events on Linux systems.
Every Linux developer, DevOps, or system admin needs to know about BPF. Its a modern day addition to tcpdump, strace, perf, ftrace and similar tools and functionality. But faster and safer. If you use those tools you'll want to use BPF tools, too.
This presentation will be by Kevin Dankwardt
The talk will be delivered via WebEx. A link a day or so before the talk will be sent to those that RSVP.
Getting Started With BPF