On Wed, Nov 18, 2020 at 04:26:52PM +0100, Alexandre Belloni wrote: > EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe > > Hello, > > On 18/11/2020 16:03:36+0100, Ludovic Desroches wrote: > > At first glance, there is no trivial way to register the pin range in the > > pinctrl-at91 driver. There is one driver for the pinctrl and one for the gpio. > > I am open to suggestions to fix it in the pinctrl-at91 driver as well if there > > is an elegant way (I have some in mind, but there are not) without having to > > refactor the driver. > > > > But shouldn't that driver be refactored at some point anyway? I know you > are moving away with new SoCs but it causes real issues. For example, > gpio hogs are not working, this is impacting some of your customers. > I agree, maintainance of this driver is difficult because of its design. Unfortunately, I doubt being able to hadnle a refactoring of this driver in a near future. > The other thing is the weird probe order preventing a nice cleanup of > the platform code. True. IMO, having gpio controlers probed before pinctrl is one of the reason which prevents a trivial fix. Regards Ludovic > > > -- > Alexandre Belloni, Bootlin > Embedded Linux and Kernel engineering > https://bootlin.com
