On 04/09, David Lechner wrote:
> On Tue, Apr 9, 2024 at 11:09 AM Marcelo Schmitt
> <marcelo.schmitt1@xxxxxxxxx> wrote:
> >
> > On 04/08, David Lechner wrote:
> > > On Mon, Apr 8, 2024 at 9:32 AM Marcelo Schmitt
> > > <marcelo.schmitt@xxxxxxxxxx> wrote:
> > > >
>
...
>
>
> > > > +static void ad4000_config(struct ad4000_state *st)
> > > > +{
> > > > + unsigned int reg_val;
> > > > + int ret;
> > > > +
> > > > + reg_val = FIELD_PREP(AD4000_TURBO, 1);
> > >
> > > Since the driver in it's current state can get anywhere near the max
> > > sample rate of ~1MSPS, I don't think it makes sense to enable turbo at
> > > this point.
> > >
> >
> > This is just enabling turbo at start up. If not enabling turbo during probe,
> > we would want(need?) to provide some interface for that, which might not be
> > much desired.
> >
>
> TURBO is only needed to achieve the max sample rate of 500k/1M/2MSPS
> on the various chips by skipping powering down some circuitry between
> samples. We can't get anywhere close to that in Linux without some
> sort of SPI offloading. So, for now, we might as well leave it
> disabled and save some power.
>
Humm, ad4000 datasheets don't really mention power usage differences for turbo
mode like ad7944 datasheet do. Though, they should be similar.
Yeah, will leave turbo disabled until providing offload support.
>
> > > > +
> > > > + st->pin_gain = AD4000_1_GAIN;
> > > > + if (device_property_present(&spi->dev, "adi,gain-milli")) {
> > > > + u32 val;
> > >
> > > Should it be an error if adi,gain-milli is set on non-adaq chips?
> >
> > Maybe. We should not change the scale if it's a chip that don't have the
> > amplifier in front of the ADC. I think the best handling would be to just
> > ignore adi,gain-milli if it's not an ADAQ device. Maybe better add a DT
> > constraint,
> > - if:
> > properties:
> > compatible:
> > contains:
> > enum:
> > - adi,adaq4001
> > - adi,adaq4003
> > then:
> > properties:
> > adi,gain-milli: false
> > ?
>
> I think this is missing a not:, but otherwise yes this should be in
> the DT bindings.
Oops, yeap, was missing the not:.
>
> Even with that though, I would still be helpful to readers of the
> driver to at least have a comment here pointing out that this property
> and related gain scaling only applies to ADAQ chips.
ok, will add a comment.
