[ath9k-devel] DMA issues with ar9280 cards

[Bernhard Schmidt <bschmidt@techwires.net>]

Hi,

[resend, due to ehlo/MX being checked case sensitive]

I have some issues using any of my AR9280 cards, identified as:
- Ubiquiti SR71E
[  172.321483] ieee80211 phy0: Atheros AR9280 Rev:2 mem=0xf9760000, irq=11
- Sparklan WPEA-110N
[   58.456404] ieee80211 phy0: Atheros AR9280 Rev:2 mem=0xf8760000, irq=11

On a recent wireless-testing kernel:
# uname -a
% Linux meshnode 2.6.38-rc6-wl-12466-g09f3227 #23 SMP Wed Feb 23 08:48:45 CET 2011 i686 GNU/Linux

I see this behaviour:
# hostapd /etc/hostapd.conf
% scan due to ht40 
[  289.389379] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  289.397106] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
[  289.548409] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  289.700270] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  289.707992] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
[  289.728509] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  289.880228] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  289.887952] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
[  289.908455] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  290.060168] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  290.067926] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
..
% scan done, setting up the BSS
[  296.060536] ath: Failed to stop TX DMA in 100 msec after killing last frame
[  296.075985] ath: Failed to stop TX DMA in 100 msec after killing last frame
[  296.083032] ath: Failed to stop TX DMA!
[  296.099187] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  296.106884] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
[  296.127247] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  297.163998] ath: Failed to stop TX DMA in 100 msec after killing last frame
[  297.182977] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  297.190677] ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
[  297.211004] ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
[  298.247659] ath: Failed to stop TX DMA in 100 msec after killing last frame
% ..

A monitor mode vif on another interface tells me that on average only
6 beacons leave the device before it's going deaf. Restarting hostapd
will make it send another 6 beacons. Behaviour in sta mode is similar
unusable, if I'm lucky enough to get any scan results the connect fails
or it goes deaf after the first few packets have been transmitted.

I tried a few different kernel version back to 2.6.36.4 which all have
the same behaviour (except the error messages, which have been mask
till lately). Also, this issue seems to be only seen on AR9820, all
other chips I have (9160, 9220, 93xx) work as expected.

Felix already told me that this is a known issue and going over other
reports on this list and linux-wireless-ml it indeed is. Just wanted
to know if there is any "known-good" kernel version or a workaround
for that?

-- 
Bernhard