On Tue, May 27, 2014 at 04:18:29PM +0200, Linus Walleij wrote:

On Mon, May 26, 2014 at 9:47 AM, Hans de Goede [off-list ref] wrote:

quoted

With level triggered interrupt mask / unmask will get called for each
interrupt, doing the somewhat expensive mux setting on each unmask thus is
not a good idea. Instead move it to the set_type callback, which is typically
done only once for each irq.

Signed-off-by: Hans de Goede <redacted>

Yes move it out of mask/unmask but no, not into set_type().

Can you not use the irqchip startup()/shutdown() callbacks
instead?

I think we can use irq_request_resources then
https://lkml.org/lkml/2014/3/12/307

We could even merge the gpio_to_irq code into it.

It's called in __setup_irq, so it's guaranteed to be called, and it
will bail out early without doing anything, since it's one of the
first thing __setup_irq does.

And how come we have no clean separation between gpiochip
and the pinmux parts here, why cant we just call
pinctrl_request_gpio() and pinctrl_free_gpio() here instead?
Or maybe pinctrl_gpio_direction_input() and have that set
up the muxing in the pinctrl driver side?

Because the function it has to be muxed to is neither gpio_in or
gpio_out, and it's not even considered a gpio. It really is just
another muxing function, like i2c or mmc.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com