Threat Model

An important question with any security project is to define the threat model:

What are we trying to protect from and what are we not trying to protect.

This project aims to protect a userspace process and its eBPF programs from anything a privileged user can do to interrupt it. Examples include:

• Signals that would kill the userspace program
• Ptrace on the userspace program
• Unauthorized eBPF map access and manipulation
• Deletion of pinned eBPF programs and eBPF maps
• Trusted loading of kernel modules*