Re: new PM branch available

[Kevin Hilman <khilman@deeprootsystems.com>]

Nicholas Chen <nchen@cs.umd.edu> writes:

> I recently checked out the latest PM branch from your git repository
> and put it on the Beagleboard (B5). I have been largely able to
> reproduce the results others have reported on this PM thread, 

...not sure what "this PM thread" you are referring to?

> but I believe my CORE and PER powerdomains are not hitting RET or
> OFF.

> I think this based on the following:
> after I
> echo 1 > enable_off_mode
> echo 1 > sleep_while_idle
> echo mem > state

CORE and PER are still on because UART clocks are not being cut in
idle:

  # echo 1 > /sys/power/clocks_off_when_idle

And see if that works.  Also at least on my Beagle, I am not able
to resume from full-chip OFF, so for starters I suggest just testing
retention: echo 0 > /sys/power/enable_off_mode

> Suspending starts but upon resume I see the following:
> ...
> Powerdomain (core_pwrdm) didn't enter target state 0
> Powerdomain (per_pwrdm) didn't enter target state 0
> Could not enter target state in pm_suspend
>
> looking at debug/pm_debug/count/
> I see that only the core_pwrdm and per_pwrdm have 0's next to OFF: and
> RET:
>
> Based on the thread, the resolution seems  to be to disable USB and
> upgrade u-boot. I think I have done both. My u-boot is 2009.03-rc2
> from Steve Sakoman's tree, and I have disabled USB support in my
> Kernel build. Is there something that I am overlooking? Also, I'm
> curious if the two powerdomains linked (i.e. I need to fix PER before
> CORE will shut off or vice versa).

For your problem, they're linked in that they both contain a UART
block.  UART1,2 are in CORE and UART3 is in PER.

Kevin