Re: [PATCH] gpiolib: Add gpio_detect, gpio_debounce and gpio_alt_func features to GPIOLIB

[Ben Dooks <ben-linux@fluff.org>]

On Mon, Jun 15, 2009 at 03:04:50PM +0200, Florian Fainelli wrote:
> Le Monday 15 June 2009 14:51:46 Ben Dooks, vous avez écrit :
> > On Mon, Jun 15, 2009 at 05:15:02PM +0800, Alek Du wrote:
> > > >From 7a76916ccea4a376a260ea67fbc79ac4d958757f Mon Sep 17 00:00:00 2001
> > >
> > > From: Alek Du <alek.du@intel.com>
> > > Date: Fri, 8 May 2009 09:46:49 +0800
> > > Subject: [PATCH] gpiolib: Add gpio_detect, gpio_debounce and
> > > gpio_alt_func features to GPIOLIB
> > >
> > > Add some more functions to GPIOLIB, they are:
> > > * gpio_detect is to set GPIO interrupt triggering method (edge, level,
> > > high, low, etc.)
> 
> The name does not seem to reflect what it does, what about gpio_set_type or 
> gpio_set_int_type for instance ?

You seem to have made the same case as myself to why this call is irrelevant
to the gpiolib layer.
 
> >
> > This is the wrong way of doing it, there is as a definit set_type
> > method in the irq_chip structure which should be used to alter the
> > way the IRQ is triggered.
> 
> I would expect your architecture IRQ handler to have a set_type callback for 
> the GPIO lines capables of generating an interrupt. See how we have beeing 
> doing it for rb532 for instance: 
> http://www.linux-mips.org/git?p=linux-queue.git;a=blob;f=arch/mips/rb532/irq.c;h=f07882029a90d3d155c17b462812c2936229458c;hb=HEAD#l173

Yes, or if the GPIO driver is exporting interrupts, the relevant handler
for that chip should have the .set_type method defined.

-- 
Ben

Q:      What's a light-year?
A:      One-third less calories than a regular year.