Re: [PATCH v5 1/2] OMAP HSMMC: Adding a Flag to determine the type of Card detect

[Tony Lindgren <tony@atomide.com>]

* kishore kadiyala <kishorek.kadiyala@gmail.com> [100621 09:49]:
> On Fri, Jun 18, 2010 at 1:49 AM, Andrew Morton
> <akpm@linux-foundation.org> wrote:
> > On Thu, 17 Jun 2010 20:56:58 +0530 (IST)
> > "kishore kadiyala" <kishore.kadiyala@ti.com> wrote:
> >
> >> --- a/arch/arm/plat-omap/include/plat/mmc.h
> >> +++ b/arch/arm/plat-omap/include/plat/mmc.h
> >> @@ -43,6 +43,9 @@
> >>
> >>  #define OMAP_MMC_MAX_SLOTS   2
> >>
> >> +#define NON_GPIO             0
> >> +#define GPIO                 1
> >
> > I'm counting about seven different definitions of "GPIO" in the kernel
> > already.
> >
> > drivers/hwmon/it87.c:
> > #define GPIO    0x07
> >
> > drivers/media/dvb/dvb-usb/ec168.h:
> >        GPIO                 = 0x04,
> >
> > drivers/net/hamachi.c:
> >        GPIO=0x6E
> >
> > drivers/staging/rtl8187se/r8180_hw.h:
> > #define GPIO 0x91
> >
> > etcetera.  It's a crazy identifier to use in a header file, and
> > the chances of a miscompile-causing collision are increasing.
> >
> >
> > enum cd_type {
> >        CD_TYPE_NON_GPIO = 0,
> >        CD_TYPE_GPIO = 1,
> > };
> 
> OK , Will repost with above changes

Please also do it the other way around where you pass CD_TYPE_NON_GPIO
in the flags in the platform data. Then there's no need to patch
all the existing boards.

Tony