I’ve reached a design fork. Honestly, this whole thing, can just be a convenience wrapper around iproute2 or over netlink sockets focused towards tc or netem . But then, I realize that there are software out there with BPF (Like AF_XDP) or what if they use RDMA? The main thing about netlink is to communicate between userspace and kernel and adding support for asynchronous messaging between the two, but the alternatives skip that. Actually, does veth even support packets going through alternate data paths? Or does it need binding to a physical netdev? and if not, would it be possible to spoof netdevices? The last question might actually blow scope up by a lot, if something doesn’t exist already

But in short, to answer the question in the title, I’ll need to go through the actual codepath a bit more. Fun.