BUG: weird behaviour after killing hostapd in latest CVS.

[JuanJo Ciarlante]

On Mon, Aug 04, 2003 at 07:27:49PM -0700, Jouni Malinen wrote:
> ...
> Which kernel version are you using?
> Which station firmware version are you using?
> Did you have wlan0sta interface UP when killing hostapd? If yes, what
> happens if you first set it down and only after that kill hostapd?

Kernel 2.4.20, firmware 1.7.4 (flashed prism2.5 PCI)  forced promisc for
hostapd+assoc_ap_addr (set_rid_word 64645 1) on a P166.

Indeed ... setting down wlan0 and wlan0sta interface BEFORE killing hostapd
behaves correctly as 0.0.4 did (recall I'm using assoc_ap_addr feature).

Some side albeit useful notes: 
 Our wireless community uses a Nokia A032 AP as a master wireless bridge,
 it doesn't  have its eth interface connected at all.
  - suspiciuosly TX speed (to AP) _seems_ somewhat better, about 20%
    (informally tested with netperf) with hostapd+assoc_ap_addr instead of
    managed mode (?)
  - hostapd+assoc_ap_addr works Ok if my macaddr is declared 
    as "normal" sta ("nid" in a032's config); if I declare my macaddr as
    "bridge" in the a032 my frames don't get re-TX by the AP.

Thanks for the excellent work... regards!

-- 
--Juanjo

#  Juan Jose Ciarlante (JuanJo PGP) jjo ;at; mendoza.gov.ar              #
#  Key fingerprint = 76 60 A5 76 FD D2 53 E3  50 C7 90 20 22 8C F1 2D    #