[Click] madwifi.stripped

John Bicket jbicket at amsterdam.lcs.mit.edu
Wed Jul 14 23:36:28 EDT 2004


Hi Nels,

I responded inline.

Nels Anderson [Nels.Anderson at colorado.edu] wrote:
>Aloha,
>
>Hello again, I have another question about the madwifi.stripped driver. 
>Currently, I'm just trying for the basics and getting it to function as a
>normal ethernet interface.  Using the gen_config_station.pl script, I load up
>that config into the /click fs and everything seems to be working fine there. 
>There aren't any messages in /click/errors.  When I cat /click/assoc/ssid and
>/click/assoc/bssid, the values appear to valid ones.  However, ath0 will not
>associate with any access point.  An iwconfig ath0 shows that access point's
>address as all 0s and a blank essid.


It looks like below that you are actually successfully associating with
an access point - you just need to assign a correct ip address to the
click interface and it should work.

The driver lets click handle all the 802.11 association protocol, so iwconfig
is only really useful for setting the wireless channel and mode (a/b/g)- 
the bssid will always be straight 0's, 

>
>I used ethereal to look at the traffic and there are absolutely no packets
>across ath0.  About once per minute, the click interface gets a packet from the
>access point destined for the ethernet address of my atheros card.
>
>The latest /var/log/messages:
>
>Jul 14 16:34:48 dhcp166 kernel: chatter: auth :: OpenAuthRequester: auth 0 seq 2
>status 0
>Jul 14 16:34:48 dhcp166 sudo:     root : TTY=pts/1 ;
>PWD=/usr/src/madwifi.stripped/etc ; USER=root ; COMMAND=/bin/sh -c echo -n
>'00:0e:d7:4d:03:30' > /click/assoc/bssid
>Jul 14 16:34:48 dhcp166 sudo:     root : TTY=pts/1 ;
>PWD=/usr/src/madwifi.stripped/etc ; USER=root ; COMMAND=/bin/sh -c echo -n
>'ITS237' > /click/assoc/ssid
>Jul 14 16:34:48 dhcp166 sudo:     root : TTY=pts/1 ;
>PWD=/usr/src/madwifi.stripped/etc ; USER=root ; COMMAND=/bin/sh -c echo -n '' >
>/click/assoc/send_assoc_req
>Jul 14 16:34:48 dhcp166 kernel: chatter: assoc-req: NODS mgt assoc_req
>00:0d:88:8a:6f:0e ->  00:0e:d7:4d:03:30 (00:0e:d7:4d:03:30) 0 Mbps +0/0
>Jul 14 16:34:48 dhcp166 kernel: chatter: assoc_resp:   40 | 1008d500 000d888a
>6f0e000e d74d0330 000ed74d 0330e013
>Jul 14 16:34:48 dhcp166 kernel: chatter: assoc :: AssociationRequester: response
>00:0e:d7:4d:03:30 +27 [ ESS ] status 0 associd 49238 ( { 2 4 11 22 } 12 18 24
>36 )
>Jul 14 16:34:48 dhcp166 kernel: chatter:
>Jul 14 16:34:49 dhcp166 sudo:     root : TTY=pts/1 ;
>PWD=/usr/src/madwifi.stripped/etc ; USER=root ; COMMAND=/bin/sh -c echo -n
>'00:0e:d7:4d:03:30' > /click/wifi_encap/bssid
>Jul 14 16:36:42 dhcp166 kernel: device click left promiscuous mode
>
>Clearly something is happening here, but I'm not sure if it's going exactly as
>it ought to be.  I don't know if I'm just missing some part of how everything
>fits together.  If there's anything else I can supply that will help illuminate
>things, I'll gladly do so.  I'm going to be out of the office for a few days,
>so the response might not be completely prompt.  As always, any
>assistance/information is greatly appreciated.

What's happend is that the element assoc::AssociationRequestor send a request
to the access point 00:0e:d7:4d:03:30 and got a response back showing
that you associated successfully (status == 0).

If you ap is running dhcpd, you probably can run dhclient or the
equivalent after this point on the click device (not ath0).

Instead of looking like packets are going through ath0, everything will
look as if it is going through the interface "click", since click will
translate the packets it gets through the "click" interface and
encap/decap them to/from the ath0 interface.



>
>-Nels
>
>P.S.  The associate.pl script doesn't handle essid's that have white space in
>the name very well.  One of ours is "Computer Science" and the script only
>comes up with "Computer".  However, there's another essid that I've been trying
>to work with that is just a single string and it's still having acting I
>described above.

yeah, sorry about that -  the script kinda sucks right now. I will probably get 
it to work, at some point, but it's a low priority. If you fix it, send me a 
patch and I will commit it.  

>_______________________________________________
>click mailing list
>click at amsterdam.lcs.mit.edu
>https://amsterdam.lcs.mit.edu/mailman/listinfo/click


More information about the click mailing list