On Tue, Sep 15, 2020 at 03:12:02PM +0000, Gehrkens.IT GmbH | Heiko Wundram wrote:
Hey all,
I've had time over the weekend to play a bit more with the radv protocol, and I've implemented some additional functional changes based on the patch I sent last week that make (at least our) configuration/usage of radv more concise. In addition to the device routes as prefixes of last week, this patch implements:
Hi Thanks for the patch. Will examine it soon and give you a feedback.
* Interface specific triggers (i.e., there can be more than one additive trigger and triggers need no longer be protocol globals, but can be specified on an interface basis) * Proper aging for routes when an exported route becomes a prefix becomes a route * Selection mechanism for which device routes to export as prefixes (the default being no export as prefix) * Sensitive prefixes (which are deconfigured and aged out normally, not only lifetime set to zero) when an interface becomes inactive/suppressed. * Route propagation per interface, not per protocol.
Generally, the configuration should be backwards compatible (i.e., same behaviour as bird 1.6.8), except for propagate routes, which now does not exist in the protocol global state; that's something that I still plan on implementing as a global toggle which can be overridden on the interface. Radv works with this patch, but is not extensively tested as the other changes I posted last week (which have been in use for somewhere around a month now).
Thanks for any feedback and hoping that somebody might find this useful!
--- Heiko.
-- Elen sila lumenn' omentielvo Ondrej 'Santiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."