Warnings/errors in use with a QCA989x card

[Michal Kazior]

On 5 June 2015 at 17:52, Ben Greear <greearb at candelatech.com> wrote:
> On 06/04/2015 10:17 PM, Michal Kazior wrote:
>> On 5 June 2015 at 03:52, Matthew Robbetts <wingfeathera at gmail.com> wrote:
>
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:08 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:09 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:09 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:09 2015] ath10k_pci 0000:02:00.0: SWBA overrun on vdev 0, skipped old beacon
>>> [Thu Jun  4 11:07:09 2015] ath10k_pci 0000:02:00.0: failed to transmit management frame via WMI: -11
>>> [Thu Jun  4 12:05:32 2015] ath10k_warn: 19 callbacks suppressed
>>> [Thu Jun  4 12:05:32 2015] ath10k_pci 0000:02:00.0: Spurious quick kickout for STA 04:db:56:e1:65:2f
>>> [Thu Jun  4 12:55:54 2015] device wlp2s0 left promiscuous mode
>>
>> This is the dreaded tx credit starvation.
>
> I have some fixes and work-arounds related to this issue in my kernels and CT firmware.
>
> Possibly it would work better for you:
>
> http://www.candelatech.com/ath10k.php
>
> The beta-14 firmware is probably best at this point..I'll promote it to official
> release soon.
>
>> In some cases if disassoc+deauth is sent and target station is asleep
>> and unresponsive it'll cause firmware to stall causing ath10k timeouts
>> during sta_state station removal. Due to insufficient credits beacons
>> can't be sent for ~10 seconds, sta_state station removal fails causing
>> mac80211 call trace splat and later spurious kickout events because
>> peer was never removed from firmware.
>>
>> There's no easy/sane fix for that, yet.
>
> I thought the newer upstream firmware had
> support for mgt frames over normal packet transport instead of over
> the WMI management interface?  Wouldn't that fix things?

I believe this was/is supposed to be available in some recent 10.2.4
builds. I think 10.2.4.48 was first supposed to support it but I see
the available "stable" fw api blob still advertises has-wmi-mgmt-tx.
The 10.2.4.48-2 from untested/ on the other hand doesn't advertise the
bit so I guess it uses the HTT path for management frames which should
prevent the tx credit starvation problem.

@Matthew: Can you try running 10.2.4.48-2 from 10.2.4/untested/ @
github.com/kvalo/ath10k-firmware and see if it helps you, please?


Michał