Re: [PATCH] s3c-adc-battery: Add naive POWER_SUPPLY_PROP_PRESENT implementation

[Anton Vorontsov <cbouatmailru@gmail.com>]

On Thu, Nov 10, 2011 at 02:05:06PM +0300, Vasily Khoruzhick wrote:
> On Thu, 2011-11-10 at 09:56 +0100, Heiko Stübner wrote:
> > Userspace programs like upower seem to depend on the presence of the
> > POWER_SUPPLY_PROP_PRESENT property.
> > 
> > The real presence of the battery cannot be sensed and all other
> > routines of s3c-adc-battery simply expect it to be present.
> > 
> > Therefore we simply return 1, unless the battery struct is not present at all.
> 
> Looks OK for me,
> 
> > Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> Acked-by: Vasily Khoruzhick <anarsoul@gmail.com>

I disagree. If the driver behaves like the hardware does not support
hot-pluggable cells, then present property is not necessary.

Upowerd should look at the 'present' only if it is there; and if it
is not, cells assumed to be always present.

[..]
> >  	if (!bat) {
> > +		if (psp == POWER_SUPPLY_PROP_PRESENT) {
> > +			val->intval = 0;
> > +			return 0;
> > +		}
> > +
> >  		dev_err(psy->dev, "%s: no battery infos ?!\n", __func__);

And in this case that dev_err() tells that either hardware does not
support hot-pluggable cells, or driver does not handle this case; which
is actually the same from the userspace point of view.

So, the problem is in upowerd, not this driver.

Thanks,

-- 
Anton Vorontsov
Email: cbouatmailru@gmail.com
_______________________________________________
linux-pm mailing list
linux-pm@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-pm