On Tue, Aug 27, 2019 at 09:42:46AM +0200, Linus Walleij wrote: > On Tue, Aug 27, 2019 at 12:00 AM Dmitry Torokhov > <dmitry.torokhov@xxxxxxxxx> wrote: > > > > Optional properties: > > > - gpios: the interrupt gpio the chip is connected to (trough the penirq pin). > > > - The penirq pin goes to low when the panel is touched. > > > + The penirq pin goes to low when the panel is touched, so make sure to tag > > > + the GPIO line with GPIO_ACTIVE_LOW. > > > > I think this is too strong. I am sure that one can come up with a way to > > connect the attention signal though polarity inverter and then one would > > have to specify GPIO_ACTIVE_HIGH in the DT. > > Yeah this is something that has been discussed recently as > Harish is suggesting to add an actual inverter to the device > tree model. As the description of reality in the device tree is > incomplete, these flags, which should nominally be consumer > flags and indicate the behavior of the consumer, becomes > a panacea for solving shortcomings in the device tree model. > > I am in favor of introducing explicit inverters in device tree, > as it makes the descriptions more consistent: e.g. regulators > are consumer oriented, they even model rails to make things > very clear on where a voltage applies, whereas GPIO polarity > is defined as "trial-and-error until whatever makes the system > run" more or less. > > > Can we say: > > > > The penirq pin goes to low when the panel is touched, so GPIO line > > should normally be tagged with GPIO_ACTIVE_LOW. > > Sure, this is not the place for hashing out DT design issues, > can you fix when applying or should I send a new version? I can fix it up on my side, once I hear from DY folks that they are OK with existing DTS changes. Thanks. -- Dmitry
