Ad-Hoc scanning measurements
Christian Bünnig
masala
Thu Apr 7 01:55:03 PDT 2005
Jouni Malinen wrote:
> On Wed, Apr 06, 2005 at 02:39:31PM +0200, Christian B?nnig wrote:
>
>
>>I want to do some measurements concerning active scanning time in ad-hoc
>>mode of 802.11b cards and I am currently browsing the hostap-driver
>>source code to find suitable places for timestamp logging (i.e. places
>>where active scanning get's finally issued and a where it is finally
>>finished). But as I am not that experienced with C and the source is
>>documented quite rarely I am not sure where to exactly place such
>>timestamp loggings.
>
>
> I have not tested scanning very thoroughly in ad-hoc mode, but I would
> guess that you would be better off with another card for this kind of
> experiment since most of the scanning is done in the firmware.
But isn't there some kind of trigger in the code which tells the card to
start a scan for a given list of channels? I've found something similar
in the linux-wlan-ng code..
>
>
>>Further I'd like to change how ad-hoc nodes handle probe-requests for
>>testing alternative handoff procedures.. is there a chance to control
>>management frame handling if the card is not in AP mode? If it is, could
>>I use 'hostapd' for this or is it necessary to hack the hostap-driver code?.
>
>
> What do you mean with handoff in context of ad-hoc mode? Please note
> that IBSS mode does not require any management frames apart from Beacon
> and Probe Req/Resp, i.e., there is no requirement for Authentication
> frames and no Association, for that matter.
Yes, there is no handoff like in managed-mode but it could be necessary
for a an ad-hoc node in a multi-hop-radio-access-network to 'switch' to
another ad-hoc-cell. To know of other ad-hoc cells it needs to scan/send
probe requests. I want the other nodes not only respond with
probe-response-packets but also inform an upper layer about the incoming
probe-requests. Obviously this is not possible if the firmware handles
probe-requests completely on its own without informing the driver. Is
that the case?
More information about the Hostap
mailing list