[Click] How to write to /proc in a click element

Robert Sombrutzki sombrutz at informatik.hu-berlin.de
Mon Jul 27 17:06:42 EDT 2009


Even if you don't want to change the channel per packet basis, it makes sence 
to use annotation instead of using syscalls. I implement a access point using 
click include parts of DFS (802.11h). There, the access point (ap) uses the 
beacon to announce a channel switch. The ap has to switch the channel after 
the beacon is send, where the time to switch is zero 
(channelswicthannouncement). So you need exact channelswitch.
For the client, i use the same mechanism. Since the client has to send all 
packet on the new channel after receiving the beacon from the ap (time to 
switch=0), i use the annotation. A click-element analyse the received beacon 
and annotates the channel to the outgoing packets. If there is no outgoing 
packet in the queue, the element generates a dummy-packet with the new 
channel, which will be discard by the driver after setting the annotated 
packet, to set the channel, which is also necessary to receive packets.
I think using annotations is a more accurate approach, then using syscalls.

Best regards,
Robert

On Montag, 27. Juli 2009, you wrote:
> Javier Sánchez wrote:
> > ok change the channel at ieee80211_hardstart() when the TX queues
> > initialize no problem.
> >
> > But u can change the channel per packet basis? i think this is not
> > possible, to change the channel the drivers needs to reset the chip
> > and restart the tx queues. It is not like change the rate , or the
> > retries..  which can be made per packet basis.
> >
> > But certainly i not tried or worked at it, may be i am wrong.
>
> This  is the least of your problems, since tuning the analogue frontend
> of a wireless card can take looooong time, I would expect it somewhere
> in the order of  of several milliseconds
>
> certainly something you don't want to do on a per-packet-basis....
>
> but since the original question was: how to do that on reception of a
> dedicated control-packet: that can make perfect sense (given that these
> control packets don't happen too often...)





More information about the click mailing list