Linux Observability with BPF

Chapter 4. Tracing with BPF

In software engineering, tracing is a method to collect data for profiling and debugging. The objective is to provide useful information at runtime for future analysis. The main advantage of using BPF for tracing is that you can access almost any piece of information from the Linux kernel and your applications. BPF adds a minimum amount of overhead to the system’s performance and latency in comparison with other tracing technologies, and it doesn’t require developers to modify their applications for the only purpose of gathering data from them.

The Linux kernel provides several instrumentation capabilities that can be used in conjunction with BPF. In this chapter we talk about these different capabilities. We show you how the kernel exposes those capabilities in your operating system so that you know how to find the information available to your BPF programs.

Tracing’s end goal is to provide you with a deep understanding of any system by taking all of the available data and presenting it to you in a useful way. We’re going to talk about a few different data representations and how you can use them in different scenarios.

Beginning in this chapter, we’re going to use a powerful toolkit to write BPF programs, the BPF Compiler Collection (BCC). BCC is a set of components that makes building BPF programs more predictable. Even if you master Clang and LLVM, you won’t probably want to spend more time than necessary building the same utilities and ensuring ...