How to APT EP. 3: Owning Thread Execution with sched_ext Hashmaps
sched_ext lets custom BPF programs replace the core Linux CPU dispatcher. But when the entire attack surface of a rootkit is a 16-byte write to an already-trusted BPF hashmap, no EDR hook-monitor will ever see it coming. Here is how a custom scheduler becomes the ultimate process inventory and targeted starvation weapon.
Aug 8, 2026
15 min
58 views How to APT EP.2: Lying to the whole netns through BPF_PROG_TYPE_FLOW_DISSECTOR
Flowdiss explores how BPF_PROG_TYPE_FLOW_DISSECTOR can be abused to forge the Linux kernel's notion of packet flow identity without modifying packet contents. By manipulating the flow_keys used to derive skb->hash, a privileged eBPF program can influence downstream consumers such as RPS, GRO, ECMP, and other eBPF programs that rely on the cached flow hash. The article presents a proof of concept, validates its effects and limitations, and demonstrates how a separate BPF LSM program can conceal the attack from standard bpftool enumeration, illustrating both an underexplored attack surface and the limits of relying on user-space BPF tooling on a compromised system.
Aug 2, 2026
22 min
1,234 views How to APT EP.1: Bypassing XDP and TC: Stealthy Connection Interception via BPF SK_LOOKUP
While most offensive eBPF techniques rely on XDP or TC to intercept raw network packets, Linux 5.6+ introduced BPF_PROG_TYPE_SK_LOOKUP, a hook operating at the socket dispatch layer. By attaching a 28-line BPF program directly to a network namespace, an attacker can silently redirect incoming TCP SYN requests to a backdoor socket without modifying packets, recalculating checksums, altering iptables rules, or leaving wire-level artifacts. This post breaks down how the technique works, common implementation traps, why it evades conventional EDR checks, and how defenders can audit for it.
Jul 30, 2026
16 min
579 views Cacheghost: Executing code from the kernel page cache
A file-backed executable mapping where the page content diverges from the backing file.
Jul 30, 2026
5 min
228 views I Spent a Week Fuzzing the BPF Verifier's New Circular Number System
I spent the better part of a week reverse-engineering Meta's ~3500 line refactoring of the BPF verifier's bounds tracking, the "circular number" (or cnum) rework that landed in Linux 7.0. I built a fuzzer, ran a quarter million random tests, traced through the linked register propagation by hand, disassembled my running kernel's vmlinux to check for backports, and generally went deeper into `kernel/bpf/verifier.c` than any sane person should.
Jul 29, 2026
4 min
144 views