On Thu, Apr 29, 2021 at 10:19 PM Lars-Peter Clausen <lars@xxxxxxxxxx> wrote: > On 4/29/21 1:35 PM, Andy Shevchenko wrote: > > On Thu, Apr 29, 2021 at 12:37 PM Lars-Peter Clausen <lars@xxxxxxxxxx> wrote: > >> On 4/29/21 11:35 AM, Andy Shevchenko wrote: > >>> On Thu, Apr 29, 2021 at 11:58 AM Lars-Peter Clausen <lars@xxxxxxxxxx> wrote: > >>>> On 4/28/21 10:22 AM, Sean Nyekjaer wrote: > >>>>> Preparation commit for the next that adds hw buffered sampling > >>> ... > >>> > >>>>> + irq = of_irq_get_byname(of_node, "INT2"); > >>>> For this I'd use device_property_match_string(dev, "interrupt-names", > >>>> "INT2"). Means it won't try to map the interrupt again, and also this is > >>>> the only place where the driver directly depends on OF, everything else > >>>> already uses the device_ API. > >>> Why not platform_get_irq_byname_optional() ? > >> Because it is not a platform device :) > > Then device_property reading like this isn't really needed. > Why? Because it doesn't bring any value in this case and actually makes readers confused. ACPI doesn't have properties started with # (they are special for OF and hiding it behind device property API is not correct). So, either leave it OF, or introduce the above API (or use existing fwnode_get_irq() by index). -- With Best Regards, Andy Shevchenko