On Wed, Oct 9, 2024 at 10:27 AM Lad, Prabhakar <prabhakar.csengg@xxxxxxxxx> wrote: > On Wed, Oct 9, 2024 at 9:11 AM claudiu beznea <claudiu.beznea@xxxxxxxxx> wrote: > > All these ports are hogs to configure them as input. Removing the hog > > property make this patch work but I'm not sure this is the right approach > > (see below diff). > > > I have dropped a query [0] to GPIO maintainers to check on the correct approach. > > https://lore.kernel.org/all/CA+V-a8vxUjTWccV-wLgy5CJiFYfEMsx-f+8weCJDP6uD_dh4AA@xxxxxxxxxxxxxx/ Yeah I replied, the callbacks in struct irq_chip rzg2l_gpio_irqchip should be calling the following callbacks: /* lock/unlock as IRQ */ int gpiochip_lock_as_irq(struct gpio_chip *gc, unsigned int offset); void gpiochip_unlock_as_irq(struct gpio_chip *gc, unsigned int offset); In its .irq_request_resources and .irq_release_resources callbacks. And it currently doesn't even define these callbacks. If the driver was using the GPIOLIB_IRQCHIP and adding the irqchip in the standard way along with the gpiochip, this would not be a problem. Can you look into simply using GPIOLIB_IRQCHIP like most other drivers as well? Yours, Linus Walleij
