Re: [PATCH 1/3] pda-power: add set_charged functionaltity

[Daniel Mack <daniel@caiaq.de>]

On Mon, Jul 20, 2009 at 10:27:52PM +0400, Anton Vorontsov wrote:
> On Thu, Jul 16, 2009 at 05:44:02PM +0200, Daniel Mack wrote:
> > This adds a function to let the power supply core know that the battery
> > is fully charged now. This can be done by watching a GPIO line, for
> > example.
> > 
> > Some battery drivers might need that information to calibrate
> > themselves.
> > 
> > Signed-off-by: Daniel Mack <daniel@caiaq.de>
> > Cc: Ian Molton <spyro@f2s.com>
> > Cc: Anton Vorontsov <cbou@mail.ru>
> > Cc: Matt Reimer <mreimer@vpop.net>
> > ---
> >  drivers/power/power_supply_core.c |   16 ++++++++++++++++
> >  include/linux/power_supply.h      |    3 +++
> 
> The patch is missing pda_power.c part?

Erm, does it? What would you expect it to do in there?

> >  #if defined(CONFIG_POWER_SUPPLY) || defined(CONFIG_POWER_SUPPLY_MODULE)
> >  extern int power_supply_is_system_supplied(void);
> > +extern void power_supply_set_battery_charged(void);
> 
> I don't quite like this interface. Note that system might want
> to know if it's supplied (hence global is_system_supplied(void) call),
> but having set_battery_charged(void) interface makes little sense
> (well, it makes sense for small embedded systems that have just
> one battery and one charger)...
> 
> There is a supplied_to psy field that might help you there...

Hmm. I thought about that too, but didn't find a way to access the
single members of the pda-power device and find the right one to
access. They're registered on a string base and matched magically once
the battery monitor is probed when found on the one-wire bus. Any
proposal how an interface could look like that does that?

All I need is a way to pass this information from my platform code down
to the battery driver.

> The cleanup patch looks OK though, I'll apply it.

Thanks. What about the one adding support for the bootstrapping of the
current_accum value? I've sent that inline as a reply to the other patch
series (4/4) last week (which was probably not a good idea).

Daniel