Re: [ANNOUNCE] updated PM branch

[Koen Kooi <k.kooi@xxxxxxxxxxxxxxxxxx>]

Op 18 dec 2008, om 00:53 heeft Kevin Hilman het volgende geschreven:

> Koen Kooi wrote:
> > Op 17 dec 2008, om 19:34 heeft Kevin Hilman het volgende geschreven:
> > > Kevin Hilman <khilman@xxxxxxxxxxxxxxxxxxx> writes:
> > >
> > > > The PM branch has been consolidated and rebased against today's
> > > > linux-omap HEAD.  I've also included a few patches that were  
> > > > posted to
> > > > the list that had fallen through the cracks.
> > >
> > > I did a couple minor updates to to the pm branch, and rebased it to
> > > current HEAD.  The updates include
> > >
> > > - more patch consolidation
> > > - VDD2 patch from Rajendra
> > > - updated IVA2 reset patch which actually works :)
> > > - included the GPTIMER12 fixes for Beagle
> > >
> > > I also tested on Beagle.  Here are some results when using a minimal
> > > defconfig (attached) and using an initramfs.  I had to disable MMC  
> > > and
> > > USB as they are currently preventing the low-level chip retention:
> > >
> > > I tested this on the latest PM branch of linux-omap:
> > >
> > > - boot idle: 342 mA
> > >
> > > - screen blank: 238 mA
> > > # echo 3 > /sys/class/graphics/fb0/blank
> > >
> > > - suspend: 85 mA
> > > # echo 1 > /sys/power/clocks_off_while_idle
> > > # echo mem > /sys/power/state
> > >
> > > - sleep-while-idle: 85 mA
> > > - this same state as suspend, but happens in idle
> > > # echo 1 > /sys/power/sleep_while_idle
> > >
> > > This is with current PM branch of linux-omap.  I understand that
> > > Tomi's new DSS tree has support to turn off the DVI framer which
> > > should save even more power, but I have not tested that.
> > Try toggling GPIO 170 manually, current consumption should drop  
> > with 60-80mA.
>
> That didn't seem to change anything for me.
>
> Looking closer, drivers/video/omap/lcd_omap3beagle.c is already  
> toggling
> GPIO 170 in current linux-omap.

Ah, so I did a better job with copy/pasting together the lcd driver  
than I thought :)

>  So the ~80mA drop above between 'boot idle' and 'screen blank' is  
> probably already counting the DVI framer.

I suspected  the 80mA drop was a bit much for just putting the DSS  
system into sleep, hence my gpio suggestion :) Thanks for confirming it.

regards,

Koen

> Kevin

Attachment: PGP.sig
Description: Dit deel van het bericht is digitaal ondertekend