Re: [PATCH 1/2] regulator: s5m8767: Modify parse_dt function to parse data related to ramp

[Mark Rutland <mark.rutland@arm.com>]

On Thu, Oct 10, 2013 at 12:10:53PM +0100, Chanwoo Choi wrote:
> On 10/10/2013 07:28 PM, Mark Rutland wrote:
> > On Thu, Oct 10, 2013 at 02:41:35AM +0100, Chanwoo Choi wrote:
> >> This patch parse 'buck[2-4]_ramp_enable and buck_ramp_delay' platform data
> >> from dts file.
> > 
> > Why?
> > 
> > What do these describe? Why do we need them?
> 
> Turn on ramp control of buck2/3/4.

Ok, but why does this need to be in the dt?

> 
> > 
> >>
> >> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
> >> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
> >> ---
> >>  drivers/regulator/s5m8767.c | 16 ++++++++++++++++
> >>  1 file changed, 16 insertions(+)
> >>
> >> diff --git a/drivers/regulator/s5m8767.c b/drivers/regulator/s5m8767.c
> >> index c24448b..cb6cdb3 100644
> >> --- a/drivers/regulator/s5m8767.c
> >> +++ b/drivers/regulator/s5m8767.c
> >> @@ -640,6 +640,22 @@ static int s5m8767_pmic_dt_parse_pdata(struct platform_device *pdev,
> >>  		return -EINVAL;
> >>  	}
> >>  
> >> +	if (of_get_property(pmic_np, "s5m8767,pmic-buck2-ramp-enable", NULL))
> >> +		pdata->buck2_ramp_enable = true;
> > 
> > Please describe what this is and why we need it.
> 
> buck2/3/4_ramp_enable is used to protect the chip from the in-rush current.

Ok. Could we not always set these in the driver? Under what
circumstances do we need this?

> 
> > 
> > For reading boolean proeprties, use of_property_read_bool.
> 
> I'll fix it.
> 
> > 
> >> +
> >> +	if (of_get_property(pmic_np, "s5m8767,pmic-buck3-ramp-enable", NULL))
> >> +		pdata->buck3_ramp_enable = true;
> >> +
> >> +	if (of_get_property(pmic_np, "s5m8767,pmic-buck4-ramp-enable", NULL))
> >> +		pdata->buck4_ramp_enable = true;
> >> +
> >> +	if (pdata->buck2_ramp_enable || pdata->buck3_ramp_enable
> >> +			|| pdata->buck4_ramp_enable) {
> >> +		if (of_property_read_u32(pmic_np, "s5m8767,pmic-buck-ramp-delay",
> >> +				&pdata->buck_ramp_delay))
> >> +			pdata->buck_ramp_delay = 0;
> > 
> > Why not initialise pdata->buck_ramp_delay to 0 beforehand?
> > 
> > That way you don't need to check the return value of
> > of_property_read_u32 here.
> > 
> > I note that pdata->buck_ramp_delay is an int, not a u32. Please use a
> > u32 temporary variable.
> 
> I'll fix it
> 
> > 
> >> +	}
> >> +
> > 
> > NAK. None of these seem to be documented in mainline's
> > Documentation/devicetree/bindings/regulator/s5m8767-regulator.txt, nor
> > does documentation seem to be added here. Until there's some description
> > of what these are and why they're needed, I don't think they're
> > suitable.
> 
> I'll add description about this patch.

Cheers.

Mark.