I think gpio_add_raw_irq_handler_with_order_priority_masked may do what you want.
Since the hardware only offers one GPIO IRQ per core, any nice interface adds complexity and latency.
To be sure, I don't understand interrupts well. I've never quite grasped the flow of what hardware and software do before and after an IRQ handler runs, and how acknowledgement works.
I certainly agree with your design philosophy that a driver should be able to register its own handler without taking over a global handler.
Since the hardware only offers one GPIO IRQ per core, any nice interface adds complexity and latency.
To be sure, I don't understand interrupts well. I've never quite grasped the flow of what hardware and software do before and after an IRQ handler runs, and how acknowledgement works.
I certainly agree with your design philosophy that a driver should be able to register its own handler without taking over a global handler.
Statistics: Posted by Eric Buddington — Mon May 06, 2024 1:24 am