[quagga-dev 5427] Re: [PATCH] RFC 2328, chap 8.1:

paul at clubi.ie paul at clubi.ie
Mon Jun 2 16:46:08 BST 2008

On Mon, 2 Jun 2008, Joakim Tjernlund wrote:

> Current quagga will try to match it with a connected remote address and
> will fail.

Right. Quagga is broken on PtP at the moment. That's a general 
brokenness issue, not a backwards compatible issue. No one using 
Quagga today can have it be adjacent with an unnumbered router.

People's network topology isn't going to change just because we fix 
that. ;)

> But this is already a known method(I think). Look at the old route
> command:
> route  [-v]  [-A  family]  add [-net|-host] target [netmask Nm] [gw Gw]
>              [metric N] [mss M] [window W]  [irtt  I]  [reject]  [mod]  [dyn]
>              [reinstate] [[dev] If]
> notice the last [[dev] If]?
> Here the details of dev If:
> dev If       force  the  route to be associated with the specified device, as
>              the kernel will otherwise try to determine the device on its own
>              (by checking already existing routes and device  specifications,
>              and where the route is added to). In most  normal  networks  you
>              won’t need this.
>              If  dev  If is the last option on the command line, the word dev
>              may be omitted, as it’s the default. Otherwise the order of  the
>              route modifiers (metric - netmask - gw - dev) doesn’t matter.

Right, you can specify just a output interface. I know how that works 
for connected networks. I'm just not 100% sure how it works with 
unnumbered, PtP interfaces. ;) I guess the equivalent of:

 	route add -net dev ppp0

should work though.

(To be honest, I was hoping someone else would pitch in on this point 
- there's at least one PPP expert subscribed to -dev ;) ).

> Saw this if the OSPF RFC too:
> [4] Note that no host route is generated for, and no IP packets can
>    be addressed to, interfaces to unnumbered point-to-point networks.
>    This is regardless of such an interface's state.

> Not sure what to make of that.

> Yes, but in my suggestion you would have to make routerID a real IP 
> address iff you want "unnumbered" PtP links too. Is that too much 
> you think?

I think that's more an administrative question.

> Perhaps one could add a command to ospf/zebra that will specify 
> such an address, default would be routerID. Then all PtP links that 
> have that address will be considered "unnumbered". That way you 
> don't have to flag every PtP i/f it is numbered or unnumbered.

Yes, zebra will have to automatically catch PtP addresses somehow. I 
suspect the easiest would be for zebra to advertise any shared 
address only as a loopback address.

Paul Jakma	paul at clubi.ie	paul at jakma.org	Key ID: 64A2FF6A
The more things change, the more they stay insane.

More information about the Quagga-dev mailing list