After upgrading to app-containers/docker-24.0.5 portage produced a warning: -- ERROR: setup CONFIG_CGROUP_BPF: is not set when it should be. WARN: setup Please check to make sure these options are set correctly. Failure to do so may cause unexpected problems. -- But when checking in the kernel config, I see that this feature depends on systemd: -- │ Symbol: CGROUP_BPF [=n] │ Type : bool │ Defined at init/Kconfig:1156 │ Prompt: Support for eBPF programs attached to cgroups │ Depends on: CGROUPS [=y] && BPF_SYSCALL [=n] │ Location: │ -> General setup │ (1) -> Control Group support (CGROUPS [=y]) │ -> Support for eBPF programs attached to cgroups (CGROUP_BPF [=n]) │ Selects: SOCK_CGROUP_DATA [=y] │ Selected by [n]: │ - GENTOO_LINUX_INIT_SYSTEMD [=n] && GENTOO_LINUX [=y] && GENTOO_LINUX_UDEV [=y] -- Not sure what is the problem here. Is this a wrong warning or is docker really depending on systemd now?
1. Yes, OpenRC + Docker work fine. 2. You're misreading it - GENTOO_LINUX_INIT_SYSTEMD requires CONFIG_CGROUP_BPF, but not the other way around.
Thanks Sam for this quick answer. I was misled by the fact this feature just doesn't appear in menuconfig, for some reason. I'm using sys-kernel/gentoo-sources-6.1.41 on amd64, do you know why it doesn't appear there? -- │ --- Control Group support │ │ [ ] Favor dynamic modification latency reduction by default │ │ [*] Memory controller │ │ [*] IO controller │ │ [*] CPU controller ---> │ │ [*] PIDs controller │ │ [ ] RDMA controller │ │ [*] Freezer controller │ │ [*] HugeTLB controller │ │ [*] Cpuset controller │ │ [*] Include legacy /proc/<pid>/cpuset file │ │ [*] Device controller │ │ [*] Simple CPU accounting controller │ │ [*] Perf controller │ │ [ ] Misc resource controller │ │ [ ] Debug controller --
np! I think the issue for you is you have BPF_SYSCALL ("Enable bpf() system call") off, so it doesn't show the cgroup option either?
Wow! I think I need a good night of sleep. Thank you for your help anyway, and sorry for the noise.