Re: [PATCH V1 4/5] backlight: qcom-wled: Add support for OVP interrupt handling

[kgunda]

On 2018-05-08 22:49, Bjorn Andersson wrote:
> On Tue 08 May 05:26 PDT 2018, kgunda@xxxxxxxxxxxxxx wrote:
>
> > On 2018-05-07 22:51, Bjorn Andersson wrote:
> > > On Thu 03 May 02:57 PDT 2018, Kiran Gunda wrote:
> [..]
> > > > @@ -220,7 +255,12 @@ static int wled_module_enable(struct wled
> > > > *wled, int val)
> > > >  				WLED3_CTRL_REG_MOD_EN,
> > > >  				WLED3_CTRL_REG_MOD_EN_MASK,
> > > >  				WLED3_CTRL_REG_MOD_EN_MASK);
> > > > -	return rc;
> > > > +	if (rc < 0)
> > > > +		return rc;
> > > > +
> > > > +	schedule_delayed_work(&wled->ovp_work, WLED_SOFT_START_DLY_US);
> > >
> > > Do you really want to delay the work on disable?
> > >
> > > Wouldn't it be better to use a delay worker for the enablement and in
> > > the disable case you cancel the work and just disable_irq() directly
> > > here.
> > >
> > Sure. Will do it in the next series.
> > > But more importantly, if this is only related to auto detection, do you
> > > really want to enable/disable the ovp_irq after you have detected the
> > > string configuration?
> > >
> > Ok. This is used for the genuine OVP detection and for the auto 
> > detection as
> > well.
>
> What is the expected outcome of detecting an OVP condition, outside 
> auto
> detection?
Ok... Out side auto detection, it is used for information purpose. I 
think it is
okay to ignore enable/disable the ovp_irq after auto detection is done.
> Regards,
> Bjorn