[Sorry, first reply was to Lennart only...] Thanks. > The relevant mechanisms are implemented via eBPF, which the kernel > restricts to privileged processes, which means --user systemd will > have a hard time. I have been expecting something like that. But this is a restriction of systemd, not the kernel, right? In other words, it is possible for a privileged user to attach BPF to an unprivileged cgroup, say, using bpftool, isn't it? (I could find that out myself, but most likely not the next one:) Assuming that it is possible kernel-wise, what is systemd's take on attaching "non-systemd" BPF to some unprivileged cgroup that it manages? Will it consider that "trampling on its toes"? Jens