[Guy Harris <guy@xxxxxxxxxxxx>]

On Friday, August 29, 2003, at 5:33 PM, David Frascone wrote:

> > So I presume that either
> >
> > 	1) LWAPP traffic goes only between access points and access routers
> > neither of which swap the word, or between access points and access
> > routers both of which swap the word;
> Exactly.
>
> > 	2) there would be interoperability problems between access points
> > 	and access routers where one swaps the word and the other doesn't;
> Correct.
>
> > 	3) there's some form of magic done by access points and/or access
> > routers to cope with that "feature".
> 1 and 3 are mostly correct.
Well, all three of those can't be correct - if 1) is true, then 2) 
isn't true because the sort of traffic that has a problem doesn't exist 
(i.e., if 1) is true, there is no LWAPP traffic between a machine that 
byte-swaps the word and a machine that doesn't), and 3) needn't be true 
as there's no need for the magic.
If 2) and/or 3) is true, then there will be bogus encapsulated 802.11 
frames (i.e., frames with the bytes swapped) sent over LWAPP - but are 
there any *other* places where bogus frames are sent?  If not, perhaps 
the "fix the bogosity" preference should be an LWAPP preference, so 
that its setting doesn't affect the dissection of regular 802.11 
packets.