power management on the 8686 chipset

[Dan Williams]

On Wed, 2009-08-26 at 19:52 +0800, Dongas wrote:
> Hi Bing,
> 
> 2009/8/26, Bing Zhao <bzhao at marvell.com>:
> > Hi Andrey,
> >
> > > -----Original Message-----
> > > From: libertas-dev-bounces at lists.infradead.org [mailto:libertas-dev-bounces at lists.infradead.org] On
> > > Behalf Of Andrey Yurovsky
> > > Sent: Tuesday, August 25, 2009 9:44 AM
> > > To: Jeff Sutherland
> > > Cc: libertas-dev at lists.infradead.org
> > > Subject: Re: power management on the 8686 chipset
> > >
> > > On Tue, Aug 25, 2009 at 9:34 AM, Jeff Sutherland<jeffs at fomsystems.com> wrote:
> > > > Can someone enlighten me as to why there seems to be such a significant
> > > > difference between the Marvell driver for the SD8686 and the libertas driver?
> > > > Or am I just out of date in working with a 2.6.27 kernel?  I need to get
> > > > power management going, especially would like to experiment with IEEE power
> > > > saving and background scanning, plus be able to put the 8686 into deep sleep
> > > > mode via some software mechanism.  (I can't simply power down the module as
> > > > that would kill the bluetooth part that I need to keep alive when out of wifi
> > > > range).  Private ioctls aren't implemented in the 2.6.27 version as in the
> > > > Marvell driver, and iwconfig seems unable to set power modes.  What is it
> > > > that I am missing here?
> > >
> > > Hi Jeff.  IEEE PS was working for me last time I tried it with 2.6.30
> > > on both the SDIO and GSPI variants.  Deep Sleep mode is not
> > > implemented in the libertas driver but there's no reason for why it
> > > couldn't be added.  Please use compat-wireless or a more recent kernel
> > > and confirm that IEEE PS is working (or not).
> > >
> >
> > The Deep Sleep support will be added to libertas driver soon.
> 
> I'm also interested in implement deep sleep for Libertas.
> One issue is that from the firmware spec 5.1 in page 90, it seems the
> host needs to sends a signal through the GPIO to wakeup the WLAN
> subsystem and this is HW dependent.
> I wonder except for the GPIO wakeup method, is there a SW method to do it?

It does not appear that there is a SW method to do it.  But isn't that
what Deep Sleep means?  If the firmware is sleeping, what is going to
bring it out of deep sleep except for a hardware interrupt triggered by
a GPIO?  Sleeping firmware can't execute commands...

Dan

> Thanks
> 
> Regards
> Benson
> 
> 
> > Regards,
> >
> > Bing
> >
> > > Thanks,
> > >
> > >   -Andrey
> > >
> > > _______________________________________________
> > > libertas-dev mailing list
> > > libertas-dev at lists.infradead.org
> > > http://lists.infradead.org/mailman/listinfo/libertas-dev
> >
> > _______________________________________________
> > libertas-dev mailing list
> > libertas-dev at lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/libertas-dev
> >
> 
> _______________________________________________
> libertas-dev mailing list
> libertas-dev at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/libertas-dev