Debugging Stalls with ath(4)

[Andrew]

Hi,

I just installed a TP-Link TL-WN851ND PCI wireless card into my 9.1-STABLE (r244942) box. It shows up in dmesg as Atheros 9287 mac 384.2 RF5133.

I am running it in hostap mode, 11g. Things seem to work OK but there are occasionally stalls in traffic across the network. e.g.

PING 10.0.0.1 (10.0.0.1): 56 data bytes
64 bytes from 10.0.0.1: icmp_seq=0 ttl=64 time=13.134 ms
64 bytes from 10.0.0.1: icmp_seq=1 ttl=64 time=77.776 ms
64 bytes from 10.0.0.1: icmp_seq=2 ttl=64 time=246.675 ms
Request timeout for icmp_seq 3
64 bytes from 10.0.0.1: icmp_seq=3 ttl=64 time=1037.327 ms
64 bytes from 10.0.0.1: icmp_seq=4 ttl=64 time=37.481 ms

--- 10.0.0.1 ping statistics ---
5 packets transmitted, 5 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 13.134/282.479/1037.327/386.148 ms

I was wondering how I would go about debugging the issue?

I have been playing with wlandebug and I can see that the client I am testing with (though it doesn't appear to be client specific) is listed as going in and out of PS mode once or twice a second. Not sure that that is a problem. Using "wlandebug rate" I see:

Jan  9 00:45:25 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 1600 (1167 bytes) OK rate/try 48 Mb /0/1
Jan  9 00:45:25 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 1600 (1554 bytes) OK rate/try 48 Mb /0/1
Jan  9 00:45:25 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 1600 (1554 bytes) OK rate/try 48 Mb /0/1
Jan  9 00:45:25 sanna kernel: wlan0: [7c:c5:37:6d:4c:7e] ath_rate_tx_complete: size 250 (30 bytes) FAIL rate/try 1 Mb /0/7
Jan  9 00:45:25 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 250 (106 bytes) OK rate/try 11 Mb /0/1
Jan  9 00:45:25 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 250 (122 bytes) OK rate/try 11 Mb /0/1
Jan  9 00:45:26 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 250 (138 bytes) OK rate/try 11 Mb /0/1
Jan  9 00:45:26 sanna kernel: wlan0: [00:23:14:98:15:b4] ath_rate_findrate: size 250 switch rate 48 (1129/690) -> 54 (736/682) after 2 packets mrr 0
Jan  9 00:45:26 sanna kernel: wlan0: [64:70:02:f0:c8:03] ath_rate_tx_complete: size 250 FAIL rate/try 0/5 no rates yet
Jan  9 00:45:26 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 250 (94 bytes) OK rate/try 11 Mb /0/1
Jan  9 00:45:26 sanna last message repeated 2 times
Jan  9 00:45:26 sanna kernel: wlan0: [00:23:14:98:15:b4] ath_rate_tx_complete: size 250 (163 bytes) OK rate/try 54 Mb /0/2
Jan  9 00:45:26 sanna kernel: wlan0: [00:17:f2:e9:04:b0] ath_rate_tx_complete: size 250 (137 bytes) OK rate/try 11 Mb /0/1

Given the proximity of the clients and the FreeBSD box I would have expected it to not have had any trouble maintaing a high connection rate - but perhaps this is a symptom of the stalls.

Can anyone suggest where to go from here?

Thanks,

Andrew