Re: kern/167491: [ath] TID != hardware queue TID in ath_tx_aggr_comp_aggr()

2012-09-09 Thread Alexander Kühn
With these in the config I see these messages: Sep 9 18:39:45 cakebox kernel: ath0: stuck beacon; resetting (bmiss count 4) Sep 9 18:42:57 cakebox kernel: ath0: stuck beacon; resetting (bmiss count 4) Sep 9 18:45:49 cakebox kernel: ath0: stuck beacon; resetting (bmiss count 4) Sep 9 18:46:28 c

Re: AR5416/Soekris instability

2012-09-09 Thread Adrian Chadd
Hi Alex, Would you please provide me with some more information about your running system? * what the hardware is (I can see in the dmesg, but I'd like to know the model of things if possible); * What (if any) works - kernel configuration, SVN revision; what does work; * What doesn't work. I'd a

Re: AR5416/Soekris instability

2012-09-09 Thread Alexander Kühn
There was no crashdump to provide - so I didn't. I had a log of a period with no traffic as well but I lost it due to the crash. I'm recompiling the kernel with the debug options now. Will retest tomorrow. This is a Soekris net5501 with a TP-Link TL-WN951N W-LAN PCI Adapter 300 MBit. I haven't tes

Re: AR5416/Soekris instability

2012-09-09 Thread Adrian Chadd
Hi, Its quite possible your environment is very noisy. Im trying to establish that. Others have reported issues on the soekris platform with ule and preemption. Try the 4bsd scheduler with preemption disabled. It wont fix the stuck beacon issues but it may stop the whole device from hanging.. Ad