The RFC patch [1] which introduces scheduler BPF into the Linux has been posted to the related mailing list (bpf and linux-kernel) two mounths ago. However, it has since only got a few comments [1][2].

To me, as a kernel-newbie, it is an interesting patch, as it enables profiling the scheduler with small overhead, changing the scheduler decision, and has barely observable overhead (NOP instruction) when no BPF programs are loaded.

Does anyone know why the discussion is so quiet?

Thanks!

[1] https://lore.kernel.org/bpf/20210916162451.709260-1-guro@fb.com/ [2] https://lwn.net/Articles/873244/