Hi, On Thu, Aug 11, 2011 at 04:07:18PM +0100, Russell King - ARM Linux wrote: > On Thu, Aug 11, 2011 at 05:15:31PM +0300, Felipe Balbi wrote: > > On Tue, Aug 09, 2011 at 09:08:01AM +0100, Russell King - ARM Linux wrote: > > > diff --git a/arch/arm/include/asm/gpio.h b/arch/arm/include/asm/gpio.h > > > index 166a7a3..15e8970 100644 > > > --- a/arch/arm/include/asm/gpio.h > > > +++ b/arch/arm/include/asm/gpio.h > > > @@ -4,4 +4,14 @@ > > > /* not all ARM platforms necessarily support this API ... */ > > > #include <mach/gpio.h> > > > > > > +#ifdef __ARM_GPIOLIB_TRIVIAL > > > +/* Note: this may rely upon the value of ARCH_NR_GPIOS set in mach/gpio.h */ > > > +#include <asm-generic/gpio.h> > > > + > > > +/* The trivial gpiolib dispatchers */ > > > +#define gpio_get_value __gpio_get_value > > > +#define gpio_set_value __gpio_set_value > > > +#define gpio_cansleep __gpio_cansleep > > > +#endif > > > > could that be a selectable symbol ? Something like: > > > > CONFIG_HAS_ARM_TRIVIAL_GPIO > > > > then on Kconfig you just: > > > > select HAS_ARM_TIVIAL_GPIO or something ? > > That makes things more complicated, because that involves digging through > a lot of platform code in a couple of places to work out exactly when we > need this - and it crosses the boundary to arch/sh too. > > So I'd prefer to keep this simple. > > The long-term goal is to remove that symbol entirely, but in order to do > that we need to kill of the "optimized" on-board SoC stuff in those (few) > gpio.h which don't have the symbol selected. This is rather necessary to > progress towards the consolidated kernel. (Re-inventing gpiolib by moving > them out of line isn't a good idea...) Ok, I understand. -- balbi
Attachment:
signature.asc
Description: Digital signature
