Problem with chillispot and dhcp relay

[Giosue Pacifico]

Hi,
I'm missing bites of conversation, is it normal?
Could the duplicate packet be solved like an arp flux? Something like
"echo 1 > /proc/sys/net/ipv4/conf/all/arp_filter".
Ciao
Giosuè

On 9/8/07, David Bird <mem.corruption at gmail.com> wrote:
> That's great. Thanks for testing! I'm happy to have chilli work with
> more environments and now supports a DHCP relay. With regard to the
> ICMP duplicate packets... I think it could be the fact that the OS is
> replying to the ping from both the dhcpif _and_ the tun interfaces.
> With your installation, are you running the supplied configuration
> with the ip up/down scripts (which setup your firewall for chilli)?
> I'll try to reproduce the problem...
>
> Thanks again.
> David
>
> On 9/8/07, Lai Zit Seng <lzs at pobox.com> wrote:
> > Ok, was too eager to try the svn update so I came back to test anyway.
> > Your latest update works great! Thanks. Especially since I imagine my
> > DHCP relay configuration is rather unusual :)
> >
> > May I bug you with one more item. This is not show stopper but
> > nevertheless something that bothers me. I can ping across chillispot
> > just fine. But if the clients ping chillispot IP itself, they get
> > duplicate replies. I.e. 1 echo-request, 2 echo-replies come back. This
> > is the case for both wired and wireless clients (so it is not a
> > quirkness of the WLC in my environment).
> >
> > The tcpdump at the client, chillispot's tun interface and chillispot's
> > dhcpif interface shows the duplicate echo-replies. The tun interface
> > capture does show something odd: There is actually an echo-reply going
> > out first, then the echo-request in, then echo-reply out. I can send the
> >
> > I can send you the tcpdump if needed :)
> >
> > Regards,
> >
> > .lzs
> >
> > ps: Copying back to mailing list.
> >
> > On 9/8/07 1:30 AM, David Bird wrote:
> > > Thanks,
> > >
> > >  From a quick review, I think my (and the current svn) version does
> > > roughly the same thing. Though, I think mine is a little cleaner in that
> > > I moved the call to dhcp_getreq() up above the newconn() in the
> > > receive_ip method. And, it returns out of receive_ip -- since the
> > > request has actually already been replied to. The newconn() method is
> > > also called from within dhcp_getreq() - and I instantiate the connection
> > > with the chaddr in the event of a relay. I look forward to hearing how
> > > the current snapshot does in your environment.
> > >
> > > David
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: chilli-unsubscribe at coova.org
> > For additional commands, e-mail: chilli-help at coova.org
> > Wiki: http://coova.org/wiki/index.php/CoovaChilli
> > Forum: http://coova.org/phpBB3/viewforum.php?f=4
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: chilli-unsubscribe at coova.org
> For additional commands, e-mail: chilli-help at coova.org
> Wiki: http://coova.org/wiki/index.php/CoovaChilli
> Forum: http://coova.org/phpBB3/viewforum.php?f=4
>
>