On Fri, 21 Sep 2018 17:34:21 +0200
Tomasz Duszynski <tduszyns@xxxxxxxxx> wrote:
> On Thu, Sep 20, 2018 at 09:37:21PM -0400, r yang wrote:
> > On Wed, Sep 19, 2018 at 09:22:39PM +0200, Tomasz Duszynski wrote:
> > > Comments inline.
> > >
> > > On Sat, Sep 15, 2018 at 01:42:14PM -0400, ryang wrote:
> > > > Add device tree support for ROHM BH1750 series ambient light sensors.
> > > >
> > > > Signed-off-by: ryang <decatf@xxxxxxxxx>
> > > > ---
> > > > drivers/iio/light/bh1750.c | 10 ++++++++++
> > > > 1 file changed, 10 insertions(+)
> > > >
> > > > diff --git a/drivers/iio/light/bh1750.c b/drivers/iio/light/bh1750.c
> > > > index a814828e69f5..50b599abb383 100644
> > > > --- a/drivers/iio/light/bh1750.c
> > > > +++ b/drivers/iio/light/bh1750.c
> > > > @@ -315,9 +315,19 @@ static const struct i2c_device_id bh1750_id[] = {
> > > > };
> > > > MODULE_DEVICE_TABLE(i2c, bh1750_id);
> > > >
> > > > +static const struct of_device_id bh1750_of_match[] = {
> > > > + { .compatible = "rohm,bh1710", },
> > > > + { .compatible = "rohm,bh1715", },
> > > > + { .compatible = "rohm,bh1721", },
> > > > + { .compatible = "rohm,bh1750", },
> > > > + { .compatible = "rohm,bh1751", },
> > > > + { }
> > > > +};
> > > > +
> > >
> > > Hmm, I believe this should be exported additionally using
> > > MODULE_DEVICE_TABLE(of, bh1750_of_match). Otherwise module loading
> > > will not work as expected.
> > >
> > > Moreover, we can do without DT and if you want to avoid
> > > warning about unused variables you may consider enclosing everything
> > > with:
> > >
> > > #ifdef CONFIG_OF
> > > ...
> > > #endif
Don't do that. There is a significant oddity now around this in that
it is possible to use a magic ACPI device ID to probe via a devicetree
binding (without CONFIG_OF). It's seriously odd, but to allow that to
work we need to not protect things with #ifdef CONFIG_OF.
(and this particular bit of fun strikes me as crazy too, but it's what
we have!)
https://www.kernel.org/doc/Documentation/acpi/enumeration.txt
(towards the bottom)
Having MODULE_DEVICE_TABLE is a good addition.
> >
> > The reason for this DT patch is because I've ported the Samsung Galaxy
> > Tab 10.1 to the mainline kernel. The device is working on the Tegra
> > platform with the right DTS file. Some peripherals on this device have
> > existing drivers such as this one which only need the DT bindings for
> > it be usuable.
> >
>
> Adding support for DT is fine. Just pointed a figner at two things
> that attracted my attention.
>
> > >
> > > > static struct i2c_driver bh1750_driver = {
> > > > .driver = {
> > > > .name = "bh1750",
> > > > + .of_match_table = of_match_ptr(bh1750_of_match),
> > > > .pm = BH1750_PM_OPS,
> > >
> > > I've recently simplified setting pm ops so this would not apply to
> > > iio/togreg cleanly.
> > >
> >
> > Does this mean I should wait until that is merged before submitting a V2
> > patch?
> >
>
> No need to wait. I guess you should rebase your work on iio/togreg
> branch and post v2.
That would be perfect as you are rerolling anyway. Let's me be lazy
and not do the merge ;)
Thanks,
Jonathan
>
> > > > },
> > > > .probe = bh1750_probe,
> > > > --
> > > > 2.17.1
> > > >
