IFF_RUNNING and SIOCGMIIPHY support

[vda]

> ----- Original Message -----
> From: "Mark Glines" <mark-hostap at glines.org>
> To: "shogunx" <shogunx at sleekfreak.ath.cx>
> Cc: <hostap at shmoo.com>
> Sent: Wednesday, March 03, 2004 6:44 AM
> Subject: Re: IFF_RUNNING and SIOCGMIIPHY support
>
> > On Tue, Mar 02, 2004 at 11:16:48AM -0500, shogunx wrote:
> > > Lets say we have a three node ad-hoc network.  10.0.0.1 can physically see
> > > 10.0.0.2.  10.0.0.3 can physically see 10.0.0.2.  10.0.0.1 and 10.0.0.3
> > > cannot physically see each other.  10.0.0.1 has another interface,
> > > 205.10.10.1, connected to the public internet, and is configured to route
> > > ipv4 packets via NAT and as an ipv6 router.  Can the
> > > traffic flow to 10.0.0.3 from 10.0.0.1, or is additional routing information
> > > required to be passed to the kernel on 10.0.0.2 and 10.0.0.3?
> >
> > Additional routing information is required.  Point-to-multipoint mode
> > OSPF would work in this scenerio.

On Wednesday 03 March 2004 08:47, Patrick Tran wrote:
> Hi ,
> As far as I understand, the scenario should be some thing similar to this
>
> 205.10.10.1<--NAT--> 10.0.0.1<---adhoc--->10.0.0.2<--adhoc-->10.0.0.3
>
> The normal ad-hoc routing 802.11 only allows nodes to communicate directly
> to each other.
> So in order to transfer traffic from 10.0.0.1 to 10.0.0.3, this ad-hoc must
> employ some routing protocol like AODV (reactive) or OLSR (proactive).
> These two are still in development process and some release working quite
> well. Patrick

Don't you trying to kill a fly with a naval cannon?

It's weird you 'stretched' 10.0.0.x network across two physical subnets.
I'd stay away from such weird configs. You can split 10.0.0.x into
two smallish 4-ip nets.

But if you did succeed in making adjacent hosts pinging themselves, then

On 10.0.0.3:
ip r a 0/0 via 10.0.0.2

On 10.0.0.2:
ip r a 0/0 via 10.0.0.1

On 10.0.0.1:
ip r a 10.0.0.3 via 10.0.0.2
ip r a 0/0 via 205.10.10.1

will make routing work.
-- 
vda