Messages in this thread |  | | | Subject | RE: [PATCH v3] mfd: ADP5520 Multifunction LCD Backlight and KeypadInput Device Driver | | Date | Tue, 6 Oct 2009 15:32:56 +0100 | | From | "Hennerich, Michael" <> |
| |
>From: Mark Brown [mailto:broonie@opensource.wolfsonmicro.com] >On Tue, Oct 06, 2009 at 01:55:45PM +0100, Hennerich, Michael wrote: >> >From: Mark Brown [mailto:broonie@opensource.wolfsonmicro.com] > >> >I'm not sure what the association with virtual gpios is? This is all >> >separate to gpiolib except in that it would mean that a gpio driver for >> >the device would be able to export these interrupts to its clients. > >> This is what I meant. >> So you propose having the MFD Core as well as its subdevs requesting the >> ADP5520 IRQ (client->irq) IRQF_SHARED? I think we already excluded us >> from using this option when we were asked to move to the NEW threaded >> irqs? > >No, I'm suggesting implementing an IRQ controller driver for the device >- register an irq_chip for the interrupt controller on it. Support for >doing this on I2C devices was added at pretty much the same time as the >IRQ_ONESHOT support. This gives access to all the genirq infrastrure >features rather than having to implement a custom IRQ handling stack. > >Like I say, I'm not sure what you meant when you were talking about >virtual gpios.
This is not an interrupt controller. The only ADP5520 subdev that needs to be notified is the adp5520-keys input driver, if present. Sounds like overshoot, registering a irq_chip using set_irq_chip_and_handler() and friends, for exactly one dedicated and known consumer.
-Michael
|  |