From mboxrd@z Thu Jan 1 00:00:00 1970 From: linux@arm.linux.org.uk (Russell King - ARM Linux) Date: Wed, 23 Jun 2010 20:54:56 +0100 Subject: [PATCH 6/6] [ARM] Introduce common definition for __io() In-Reply-To: <1277293938-25804-7-git-send-email-eric.y.miao@gmail.com> References: <1277293938-25804-1-git-send-email-eric.y.miao@gmail.com> <1277293938-25804-7-git-send-email-eric.y.miao@gmail.com> Message-ID: <20100623195456.GC25070@n2100.arm.linux.org.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jun 23, 2010 at 07:52:18PM +0800, Eric Miao wrote: > Signed-off-by: Eric Miao > --- > arch/arm/include/asm/io.h | 4 ++++ > arch/arm/mach-aaec2000/include/mach/io.h | 1 - > arch/arm/mach-at91/include/mach/io.h | 1 - > arch/arm/mach-bcmring/include/mach/io.h | 1 - > arch/arm/mach-clps711x/include/mach/io.h | 1 - > arch/arm/mach-cns3xxx/include/mach/io.h | 1 - > arch/arm/mach-davinci/include/mach/io.h | 1 - > arch/arm/mach-ep93xx/include/mach/io.h | 1 - Again, ebsa110 is missing. That's because __io() is a two-state definition. It means something if a platform does _not_ define it. It also means something _if_ a platform defines it. You can't move it to asm/io.h because that precludes the first state from ever occuring. It's not a simple case that those platforms can provide some subsitute definition for __io() - they have to supply their own inb() et.al. accessors because they need to do something uniquely special there.