Re: [PATCH v8 0/11] dmtimer adaptation to platform_driver

[Kevin Hilman <khilman@ti.com>]

Hi Tarun,

"DebBarma, Tarun Kanti" <tarun.kanti@ti.com> writes:

>> >
>> > Also, testing with PM on 34xx/n900, I noticed that this series prevents
>> > PER and CORE from hitting retention during suspend.  I haven't debugged
>> > why yet.
>> I have not done power testing. I will try this out right away and confirm.
>>
> Here is a brief update on pm test.
>
> (1) Without dmtimer patch series:
> [  305.641204] Powerdomain (core_pwrdm) didn't enter target state 1
> [  305.647491] Could not enter target state in pm_suspend

What board is this on?

Indeed some platforms do not hit CORE retention by default, most often
because of the boot loader leaving some devices in a state that prevents
them from idling.

At least on 34xx/n900 (with DSS enabled), core is hitting retention for
me.  On Beagle, I have to also use my pm-otg-reset branch to ensure that
the MUSB block is properly reset on boot, othewise MUSB is preventing
CORE RET.  I suspect this would be the same problem on 3430SDP.

> (2) With dmtimer patch series:
> [   25.291503] Powerdomain (core_pwrdm) didn't enter target state 1
> [   25.297790] Powerdomain (per_pwrdm) didn't enter target state 1
> [   25.303985] Could not enter target state in pm_suspend
>
> I will debug further.

Thanks, you can see from this that something in the dmtimer series is
preventing PER from hitting retention.

Kevin