[Click] nsclick wireless

[Gmail]

Eddie,

On 6/30/07, Eddie Kohler <kohler at cs.ucla.edu> wrote:
>
> Hi Nicolas, Szymon,
>
> I am not sure that this patch is ready to be accepted into mainline Click.
> For instance, I am concerned that 802.11 stuff leaks its way into the
> ClickClassifierClass.


The only 802.11 stuff that should be in there is to access the ethernet
addresses correctly. These addresses are translated to ns node id's and vice
versa. There is one exception about retry marking in packets. I should have
implemented this feauture in the mac layer, so I will move this code to
where it belongs.

  Will that work for non-802.11 nsclick configurations,
> and if so, why?


Yes, the whole extension was written with backwards compatibility in mind.
The extension auto detects and differentiates between raw wireless frames
and classic ethernet frames by checking the wifi magic header. In the former
case the 802.11 mac behaves as an atheros card , in the latter case it
behaves as a "classic" 802.11 ns mac.

So a basic scenario using 802.3 and 802.11 network will continu to work
without modification. I verified this with the nsclick-simple-wlan scenario
and it behaved as expected (unless I missed something ;-) ).

  Some of the other changes in the ns patches you supply, look
> like they are adding non-Click functionality to ns; these patches should
> be
> submitted to mainline ns.  Can you argue for this patch?


Indeed there are some small bugfixes for ns including Szymons patches. I
understand that the goal of nsclick is not to fix ns bugs, so I can remove
them from the patch set if you want. However I am afraid that sending the
whole patchset to the ns mainline is difficult as the purpose of this set is
to support the wireless feautures of the Click Modular Router.

Greetings,
Nico

Eddie
>
>
> Gmail wrote:
> >
> > Dear Eddie,
> >
> > Earlier this month Szymon Chachulski sent some nsclick improvements to
> this
> > mailinglist. In order to aggregate all nsclick patches for the upcoming
> > Click-1.6 release, we have packaged these into our ns-madwifi patch, in
> > agreement with Szymon.
> >
> > The attached patch will both enable ns-madwifi and resolve the issues
> > Szymon
> > Chachulski discovered. We have put additional documentation about
> > ns-madwifi
> > online at <http://www.pats.ua.ac.be/software/nsmadwifi>.
> >
> > Regards,
> > Nicolas Letor
> >
> > --
> > Nicolas Letor
> > University of Antwerp
> > Dept. Math. and Computer Science
> > PATS - Performance Analysis of Telecommunication Systems Research Group
> > Middelheimlaan 1,
> > B-2020 Antwerp, Belgium
>



-- 
Nicolas Letor
University of Antwerp
Dept. Math. and Computer Science
PATS - Performance Analysis of Telecommunication Systems Research Group
Middelheimlaan 1,
B-2020 Antwerp, Belgium