Skip to content

feat: make links, addresses, routes and kernel settings by netlink instead of ip - #134

Merged
leonkrix merged 2 commits into
mainfrom
feat/netlink-layer
Nov 12, 2026
Merged

leonkrix merged 2 commits into
mainfrom
feat/netlink-layer

Conversation

@leonkrix

Copy link
Copy Markdown
Owner

The Kernel backend builds links, veth pairs, addresses, fixed neighbours, routes (plain and multipath), MTU and segment limits, and writes the kernel settings of a node, by netlink instead of ip and the hidden sysctl call of the supervisor (BKD-05, D-101). The socket of a node is made by a thread that has entered the namespace with setns; no unsafe. tc is still used for queues and filters.

New dependencies: netlink-packet-core, netlink-packet-route, netlink-sys. The hidden sysctl subcommand of the supervisor is removed. D-098: the sentence about the threshold of the timer is corrected with the numbers of the runners.

@leonkrix
leonkrix merged commit 2bf1ae3 into main Nov 12, 2026
36 checks passed
@leonkrix
leonkrix deleted the feat/netlink-layer branch November 12, 2026 09:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant