On Sun, Oct 21, 2018 at 09:46:20AM +0200, Stefan Sperling wrote: > On Sun, Oct 21, 2018 at 08:25:02AM +0200, def...@posteo.de wrote: > > Hi > > > > Any updates? > > I still want to do some testing at my end with ath hardware I have. > So please just wait. >
I have now tried several AR5424 devices with your patch: ath0 at pci6 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 16 ath0: AR5424 14.2 phy 7.0 rf 10.2 eeprom 5.3, WOR5_ETSIC, address 00:22:69:78:99 :5c ath0 at pci6 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 16 ath0: AR5424 14.2 phy 7.0 rf 10.2 eeprom 5.3, WOR5_ETSIC, address 00:22:69:65:c6:ff ath0 at pci6 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 16 ath0: AR5424 14.2 phy 7.0 rf 10.2 eeprom 5.3, WOR5_ETSIC, address 00:23:4d:cd:f4:79 None of the above devices work correctly. I doubt this is a harware problem, since all 3 devices show the same behaviour. I have tested both client mode and hostap mode. Client mode can manage to associate but does not get a DHCP lease. Hostap mode does not send beacons so the AP doesn't show up in scans. tcpdump -y IEEE802_11_RADIO -i ath0 reveals that receiving frames works. But transmitting frames evidently does not work reliably. Either something is broken in the transmit code path or the radio isn't calibrated correctly: ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout ath0: device timeout Is this patch really working well for you?