Re: [PATCH 1/2] gpio: add pin biasing and drive mode to gpiolib

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

On Wed, 20 Apr 2011 14:38:01 +0200
Linus Walleij <linus.walleij@linaro.org> wrote:

> 2011/4/20 Ben Nizette <bn@niasdigital.com>:
> > On 19/04/2011, at 6:38 PM, Alan Cox wrote:
> >> Leaving aside the current input/output and on/off bits I would go for
> >> being able to do
> >>
> >>       gpio_get_property(gpio, GPIO_BIAS, GPIO_BIAS_WHATEVER);
> >>       gpio_set_property(gpio, GPIO_BIAS, GPIO_BIAS_WHATEVER_ELSE);
> >
> > Yeah I'm all for that so long as the capability constants are defined by the gpio
> > provider, eg <linux/gpio/mygpioexpander.h>.  There's no way gpiolib should be
> > keeping a big ole list of every possible config option for every gpio provider.
> 
> OK I buy that. I will refactor this solution to some opaque call instead
> and start from there.
> 
> > Well, maybe gpiolib can know about the options (eg GPIO_BIAS) so long
> > as it doesn't have to enumerate every possible value.
> 
> I will drop that even, one parameter is better than two if one of them
> is custom nevertheless. What difference does it make..

One parameter means its completely useless and we'll have to go change
the API.

Without the 'operation' being a parameter of its own no driver knows how
to answer the question 'is this shared operation A'

Alan