Re: [PATCH V2 3/3] iio: adc: Add support for PMIC7 ADC

[Jishnu Prakash]

Hi Andy,

On 4/27/2020 6:58 PM, Andy Shevchenko wrote:
> On Mon, Apr 27, 2020 at 3:56 PM Jishnu Prakash <jprakash@xxxxxxxxxxxxxx> wrote:
> > On 4/17/2020 3:51 PM, Andy Shevchenko wrote:
> > On Thu, Apr 16, 2020 at 1:48 AM Jishnu Prakash <jprakash@xxxxxxxxxxxxxx> wrote:
> Stop using HTML. It breaks badly the reply and discussion.
>
> ...
>
> > +static const struct adc5_data adc7_data_pmic;
> >
> > Global variable? Hmm...
> >
> > adc7_data_pmic is referenced twice before its actual definition (which was added along with corresponding adc5_data struct for PMIC5 ADC), so I have given the initial declaration here.
> Maybe you can realize how to avoid global variable at all?
There is a way to remove this, I'll make this change with some other 
changes in the fifth patch of my latest post.
> ...
>
> > +       buf[1] &= 0xff & ~ADC5_USR_FAST_AVG_CTL_SAMPLES_MASK;
> >
> > What the point of 0xff & part?
> >
> > This was something you suggested in my first post:
> >
> > > +       buf[1] &= (u8) ~ADC5_USR_FAST_AVG_CTL_SAMPLES_MASK;
> > Use '0xFF ^ _MASK' instead of casting.
> >
> > ...
> >
> > > +       buf[3] &= (u8) ~ADC5_USR_HW_SETTLE_DELAY_MASK;
> > Ditto.
> >
> > I think "0xff &" works as intended here in place of casting to (u8)...
> Does it work without casting? (Note, I suggested slightly different expression)
> I.o.w. what the problem casting solves?
I checked this part again. It looks like casting is not strictly 
required here, I'll remove it in my latest post.
> > +       buf[1] |= prop->avg_samples;
> > +
> > +       /* Select ADC channel */
> > +       buf[2] = prop->channel;
> > +
> > +       /* Select HW settle delay for channel */
> > +       buf[3] &= 0xff & ~ADC5_USR_HW_SETTLE_DELAY_MASK;
> >
> > Ditto.
> >
> > +       buf[3] |= prop->hw_settle_time;