Re: [RFC PATCH v3 3/5] platform/chrome: Introduce device tree hardware prober

[Chen-Yu Tsai]

On Sat, Dec 2, 2023 at 8:58 AM Doug Anderson <dianders@xxxxxxxxxxxx> wrote:
>
> Hi,
>
> On Tue, Nov 28, 2023 at 12:45 AM Chen-Yu Tsai <wenst@xxxxxxxxxxxx> wrote:
> >
> > @@ -61,6 +61,17 @@ config CHROMEOS_TBMC
> >           To compile this driver as a module, choose M here: the
> >           module will be called chromeos_tbmc.
> >
> > +config CHROMEOS_OF_HW_PROBER
> > +       bool "ChromeOS Device Tree Hardware Prober"
>
> Any reason that it can't be a module?

No technical one. However if it's a module, the user has to manually load
it. So I think it's more of a usability thing.

OOTH I think this needs to be a module if I2C is built as a module.
Somehow I had thought of it at one point but then it slipped my mind.

> > +       depends on OF
> > +       depends on I2C
> > +       select OF_DYNAMIC
> > +       default OF
>
> You probably don't want "default OF". This means that everyone will
> automatically get this new driver enabled which is unlikely to be
> right.

I thought this whole section was guarded behind KCONFIG_CHROME_PLATFORMS.
So if the user has CHROME_PLATFORMS enabled and has OF enabled, they
likely need the prober.

> > +static int chromeos_of_hw_prober_probe(struct platform_device *pdev)
> > +{
> > +       for (size_t i = 0; i < ARRAY_SIZE(hw_prober_platforms); i++)
> > +               if (of_machine_is_compatible(hw_prober_platforms[i].compatible)) {
> > +                       int ret;
> > +
> > +                       ret = hw_prober_platforms[i].prober(&pdev->dev,
> > +                                                           hw_prober_platforms[i].data);
> > +                       if (ret)
>
> Should it only check for -EPROBE_DEFER here? ...and then maybe warn
> for other cases and go through the loop? If there's some error
> enabling the touchscreen I'd still want the trackpad to probe...

Makes sense. However there's no extra information to give in the
warning though.

> > +                               return ret;
> > +               }
> > +
> > +       return 0;
>
> Random thought: once we get here, the driver is useless / just wasting
> memory. Any way to have it freed? ;-)

I don't think there is a good way to do that, except maybe marking all
the functions as __init? But that likely doesn't work in combination
with deferred probing (say the i2c driver is a module).

ChenYu