First attempt at o11s. Need hints.

Steve Brown sbrown at cortland.com
Sun Jul 27 18:30:17 EDT 2008


Steve Brown wrote:
> Steve Brown wrote:
>>
>> I have build a patched wireshark. It records exactly 11 frames from 
>> each machine. A copy is attached. Neither machine responds to the 
>> other. There is no mac80211 debug output, but some from b43, so I'm 
>> sure that kernel.debug is getting to the log.
>>
> Turns out the wireshark patch assumes a mesh header length of 5, not 
> the 6 that mac80211 thinks it is now.
>
> It's now fixed, but the o11s results are the same.
>
> Better dump attached.
>
> I tested the radios in ad-hoc mode and they can ping each other, but 
> as mesh points, nothing.
>
> It looks like the mesh ie codes changed too. If I can get further than 
> I have, I'll test the ie's and post an updated wireshark patch.
>
> Steve
>
>
Found the bad commit, "revamp beacon configuration" (7/9/08), and posted 
the problem to linux-wireless. With the reverted patch, the zd1211rw and 
the b43 find and talk to each other.

I conformed the ie codes in wireshark to those in mac80211. They seem to 
decode the data with the reverted patch. After I've tested a little 
more, I'll post the wireshark changes back here.

I'm very impressed with your work!

Steve





More information about the Devel mailing list