[Avila] latest using madwifi + hostapd

[devel]

------=_NextPart_000_004C_01C67046.D94F1CA0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hey Tim and list,

First off, my system specs:

    Board:            Avila GW2342
    Kernel:           0.6 dev kit's 2.6.15-uc0
    Mini-PCI:        Senao NL-5354MP PLUS ARIES2(F)
                            Atheros chip labeled with AR5213-A-00
    Madwifi:          Various dailysnap shots over the past month, =
currently using 1538 (May 2, 2006 I believe)
    Hostapd:         Various versions as well, been using 0.4.7, 0.4.8, =
0.5.2, & 0.5.3.

Now, to some of my findings while running all of this on the Avila =
GW2342.

  a.. Madwifi driver, all revs compile and run, standalone, on the =
Avila. Including using 128-bit WEP.
  b.. Clients associate, disassociate, come and go at will, no issues. =
This is done with mulitple client systems and client Wifi devices.
  c.. Hostapd, this is where the fun (or lack there of) begins. Some of =
the time, hostapd initiates and seg faults with the message, "Internal =
error: Oops!!!" on the console along with the core dump. I assume this =
is all from the driver, however, but I could be wrong. I can never seem =
to recover from this, a reboot is the only thing to do. Unloading the =
modules (madwifi) can not be performed.
  d.. During those times that hostapd does initiate without a seg fault, =
I have a problem with clients connecting to the AP. In the syslog and on =
the console, if I have started hostapd with -d or -dd, I see the line =
"ath0: STA xx:xx:xx:xx:xx:xx IEEE 802.11: deauthenticated due to local =
deauth request", repeatedly. All clients are "associated" with the AP, =
but not "authenticated" (hopefully you understand what I mean here).
My assumption after all of this is, I believe it to be a problem with =
the madwifi driver, since I can reproduce this problem on the same =
driver, will all the hostapd versions mentioned above. However, again, =
this is only my assumption. It could be hostapd or both. I may have not =
cross-compiled either or both correctly. I do believe I have, but I'll =
be the first to say I'm not perfect. So there it is in a flash. If =
anyone can possibly help with this or if anyone else has had similiar =
issues, lets work together to find a solution. Thanks.

Travis
------=_NextPart_000_004C_01C67046.D94F1CA0--