On Sun, Oct 31, 2021 at 2:25 PM Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote: > > On Thu, Oct 14, 2021 at 3:48 PM Jonas Gorski <jonas.gorski@xxxxxxxxx> wrote: > > > > of_property_read_u32 returns 0 on success, not true, so we need to > > invert the check to actually take over the provided ngpio value. > > > - if (of_property_read_u32(dn, "ngpios", &num_gpios)) > > + if (!of_property_read_u32(dn, "ngpios", &num_gpios)) > > chip->gc.ngpio = num_gpios; > > Just a side note: I'm wondering how many of such we have in the > drivers and why we can't simply do it under the OF GPIO library's hood > for everybody. > > -- > With Best Regards, > Andy Shevchenko Looks like we have 17, so yes, definitely a candidate for pulling into the GPIO core. Bart