Re: [PATCH] [drivers] [SPI] SPI_GPIO: add support for controllers with missing MISO pin

[Daniel Mack <daniel@caiaq.de>]

On Mon, May 18, 2009 at 09:34:12AM +0100, Ben Dooks wrote:
> > > >  static inline int getmiso(const struct spi_device *spi)
> > > >  {
> > > > -	return !!gpio_get_value(SPI_MISO_GPIO);
> > > > +	if (SPI_MISO_GPIO)
> > > > +		return !!gpio_get_value(SPI_MISO_GPIO);
> > > > +	else
> > > > +		return 0;
> > > >  }
> > > 
> > > Is zero a good approximation for 'no gpio' ?
> > 
> > Now I found that zero might be a valid gpio pin number on some architectures
> > (it just means GPIO0 pin). This is imho a bit strange behavior of gpiolib as
> > there should be also a special values for INVALID or NOGPIO cases. Does 
> > anyone have any ideas how such cases should be handled properly? 
> 
> I belive there is a gpio_is_valid() function to tell you precisely if the
> given GPIO is valid.

And then -1 can be passed in from the platform data which will fail the
gpio_is_valid() test.

Daniel