Investigating Issues in Linux Embedded Systems

This session presents methods for finding the causes of issues in embedded Linux systems, including hardware restrictions and software bugs. It demonstrates kernel debugging, memtest, ftrace, perf, and memory logging through examples involving drivers, NFS root filesystems, kernel panics, and performance.

Syllabus

Intro
Overview of this session
When do we observe issues?
What kind of issues do we observe?
How to find the root cause?
Investigation ways - #define DEBUG
Investigation ways - memtest
Investigation ways - ftrace
Investigation ways - perf
Investigation ways - memory logging example
Case studies - The USB HID driver cannot work correctly
Case studies - NFS rootfs cannot work on a new SoC
Case studies - Kernel panic while data transferring
Case studies - Probing a device driver fails
Case studies - Performance issue
Summary