"Mario Rossi" <mariofutire@xxxxxxxxxxxxxx> writes: > So far I've understood that wpa_supplicant is responsible for setting > the key for the interface. > When I debug it, I just add "iwconfig" at the end of ifup-interface > (so it prints all the parameters of the wireless interface) and I can > tell if something went wrong. > In normal situation, the key has to be set, only the the dhcp can work. I have an atheros cardbus card made by ubiquiti. Every once in a while wpa_supplicant gets in a loop where it negotiates the key, I very briefly see "completed" with the wpa-psk, CCMP, etc displayed in wpa_gui. Before I can blink the wpa_gui says DISCONNECTED and all the vitals are cleared. It then repeats the above forever. The fix is usually to do a "service wpa_supplicant restart". Sometimes I need to take down the ath0 interface and bring it up too. It feels like a timing problem where wpa_supplicant just isn't waiting long enough for something. BTW. I just spent an hour trying to get the network working with wpa_supplicant started before the network code. If I put an "exit 0" at the start of ifup-wireless then the dhclient never got started, even though wpa_gui shows the association exists. Ipv6 is also up, so the network is usable for that. I have no idea why dhclient didn't get started. The only thing I can guess is timing also. I've returned to starting wpa_supplicant from the bottom of ifup-wireless and things are back to normal. I guess I need to dig up the sources to wpa_supplicant and watch this run in gdb and see if I can find the reason for the looping error. Maybe that will shed some light on why starting order of wpa_supplicant and the network card is so delicate. -wolfgang -- Wolfgang S. Rupprecht http://www.wsrcc.com/wolfgang/ IPv6 on Fedora 7 http://www.wsrcc.com/wolfgang/fedora/ipv6-tunnel.html
