On 4.6.2021 9.35, Tony Lindgren wrote:
> Hi,
>
> * Jarkko Nikula <jarkko.nikula@xxxxxxxxxx> [210602 18:21]:
>> Hi
>>
>> On 6/2/21 8:30 AM, Tony Lindgren wrote:
>>> Sounds like the beagleboard timer errata handling is either picking
>>> a wrong clockevent timer, or later on ti-sysc misdetects the active
>>> timer for clockevent and idles it.
>>>
>>> What does dmesg say in the beginning for clockevent and clocksource
>>> timers?
>>>
>> I attached logs from commit 6cfcd5563b4f and commit 30c66fc30ee7 before
>> it. Perhaps this is relevant difference between them?
>>
>> +ti-sysc: probe of 49032000.target-module failed with error -16
>> +ti-sysc: probe of 48304000.target-module failed with error -16
>
> Yeah so it seems. We now attempt to ignore the system timer instances
> while it seems we need also block idling at least for timer12 that is
> used for the beagle timer.
>
> Can you test the following patch and see if it helps?
>
It works! I tested on top of both regressing commit and the most recent one:
6cfcd5563b4f ("clocksource/drivers/timer-ti-dm: Fix suspend and resume
for am3 and am4")
and
f88cd3fb9df2 ("Merge tag 'vfio-v5.13-rc5' of
git://github.com/awilliam/linux-vfio")
You may add:
Tested-by: Jarkko Nikula <jarkko.nikula@xxxxxxxxxx>
