spurious quick kickout and ath10k firmware crash

[Matti Laakso]

I have one Android device (Nexus 7 2013 model), which doesn't like to
play with ath10k. I often get dmesg full of spurious quick kickouts for
that device, and this time it was accompanied with a firmware crash:

...
[152937.200000] ath10k: Spurious quick kickout for STA d8:50:e6:8f:bd:ef
... continues ...
[153128.210000] ath10k: Spurious quick kickout for STA d8:50:e6:8f:bd:ef
[153129.850000] ath10k: firmware crashed!
[153129.860000] ath10k: hardware name qca988x hw2.0 version 0x4100016c
[153129.860000] ath10k: firmware version: 10.1.467.2-1
[153129.870000] ath10k: target register Dump Location: 0x0040AB94
[153129.880000] ath10k: target Register Dump
[153129.880000] ath10k: [00]: 0x4100016C 0x000015B3 0x0099D923 0x00955B31
[153129.890000] ath10k: [04]: 0x0099D923 0x00060330 0x0000001C 0x0040AD04
[153129.890000] ath10k: [08]: 0x0000000D 0x0040AE10 0x00442088 0x00955A00
[153129.900000] ath10k: [12]: 0x00000009 0x00000007 0x009AB338 0x009AB369
[153129.910000] ath10k: [16]: 0x00958080 0x00953869 0x00000000 0x00000000
[153129.910000] ath10k: [20]: 0x4099D923 0x0040ACA4 0x00000028 0x00407120
[153129.920000] ath10k: [24]: 0x8099C99C 0x0040AD04 0x00000000 0xC099D923
[153129.930000] ath10k: [28]: 0x8098EB29 0x0040AD74 0x00442088 0x0040AE08
[153129.930000] ath10k: [32]: 0x80994E4B 0x0040ADD4 0x0041ED48 0x00439EB0
[153129.940000] ath10k: [36]: 0x809968F1 0x0040ADF4 0x0041ED48 0x0042503C
[153129.950000] ath10k: [40]: 0x80942EB3 0x0040AEE4 0x00423348 0x00000002
[153129.950000] ath10k: [44]: 0x80940F18 0x0040AF14 0x00000005 0x004039E4
[153129.960000] ath10k: [48]: 0x80940EEA 0x0040AF44 0x00400000 0x00000000
[153129.970000] ath10k: [52]: 0x80940F31 0x0040AF64 0x00401C10 0x00400600
[153129.970000] ath10k: [56]: 0x40940024 0x0040AF84 0x004068E8 0x004068E8
[153130.980000] ath10k: suspend timed out - target pause event never came
[153131.500000] ieee80211 phy0: Hardware restart was requested
[153132.780000] ath10k: device successfully recovered

My ath10k is not quite up to date, it is from backports at the end of
March, up to and including ath10k: do not overwrite max_antenna_gain.
I'll update soon and see if I can reproduce that. Any other way to help
(suitable debug mask for example)?

Matti Laakso