Re: ath9k crash 3.2-rc7

 > thanks for the providing the log. it looks like there is some other
 > problem in doing the chip reset and
 >eventually when we get into ath_update_survey_stats we are hitting the
 >panic.
 > you can try two things:
 > 
 > *after these WARNINGS, is your card in some stable state?. are you
 > able to see it in lspci(ie does it gets disappeared) and still
 > continue your wifi operation. if not then there is some other problem
 > while the panic is a symptom of it.

There is another failure mode. 

I think rmmod/modprobe says all:

Jan  7 00:25:04 401a0bf1 kernel: [ 5373.686727] ath9k: Driver unloaded
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.441778] ath9k 0000:03:00.0: enabling 
device (0000 -> 0002)
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.441789] ath9k 0000:03:00.0: PCI INT A 
-> GSI 17 (level, low) -> IRQ 17
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.441802] ath9k 0000:03:00.0: setting 
latency timer to 64
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.558038] ath: Couldn't reset chip
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.558042] ath: Unable to initialize 
hardware; initialization status: -5
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.558048] ath9k 0000:03:00.0: Failed to 
initialize device
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.558092] ath9k 0000:03:00.0: PCI INT A 
disabled
Jan  7 00:25:07 401a0bf1 kernel: [ 5376.558259] ath9k: probe of 0000:03:00.0 
failed with error -5

When I had the card just stop working without crashing my notebook (3.0, 3.1) 
I got the same when trying to reload the driver.

 > *does the other patch which remove some suspicious looking code, to
 > see if it improves the situation and avoids the warnings itself.

I have built and loaded the kernel with this patch. I even have netconsole 
apparently working. But ten hours have not elapsed yet. So far so good.

If it works till the 10th, I will test it with roaming, too.