From: Manjunathan PY (pymanju_at_miel.mot.com)
Date: 2002-07-16 06:04:00 UTC
Do you have access to Cisco Airnet firmware source code?
if yes, can pls email the location where i can find the same.
From: al_at_adalinux.org [mailto:al_at_adalinux.org] Sent: Tuesday, July 16, 2002 6:58 AM
Subject: Re: Cisco Aironet 350
On Mon, Jul 15, 2002 at 09:22:22PM +0300, Jouni Malinen wrote:
> Do you mean that you use hostapd from the Host AP driver and modified
> airo_cs.o driver? I have considered testing such combination at some
> point, but haven't thus far have time for this. Anyway, I would be
> interested in getting to test your changes.
> I have been aware of the AP mode in Cisco cards for quite some time, but
> testing it has always been delayed (enough interesting things to do with
> Prism2 driver ;-). Have you confirmed that the card sends correct beacon
> frames when set into AP mode and whether it replies to probe requests?
> If one of these is not working fine, there might be some errors with
> different station card models.. Anyway, dumping all 802.11 frames with
> 3rd host in monitor mode and going through them, e.g., with Ethereal,
> could give useful information for debugging the association problems.
> Jouni Malinen PGP id EFC895FA
I am using a very recent (28th June) cvs snapshot from airo-linux (http://sourceforge.net/projects/airo-linux) on (currently) RedHat 7.2 with kernel upgrade to 2.4.18-3. Wireless tools are version 11 (the main driver file airo.c tests the version of wireless extensions). Hostapd.c came from Prism2-2002-05-19.tar.gz.
I was initially modifying the driver for reasons other than making it run as an AP but as far as I can tell from a quick scan the only relevant change I have made is to add a command (ioctl) to set the flags (FLAG_802_11) to enable capturing raw packets with minimal effect on behaviour - I have modified hostapd to use this (my copy of hostapd.c has become fairly polluted, there may be other relevant changes there).
This version (certainly) of the driver will send beacon frames and probe responses without modification if mode is set to "master".
Beacon frames and probe responses appear to be correct (monitored with a commercial tool) and eg. I have (just now) tried a Netgear Prism2 and it attempts to authenticate although it is not happy with the response.