From: linux@arm.linux.org.uk (Russell King - ARM Linux) Date: Wed, 3 Feb 2010 10:18:02 +0000 Subject: [PATCH 10/16] ARM: LPC32XX: Converted most register types to void __iomem * In-Reply-To: <1265155168-28909-11-git-send-email-wellsk40@gmail.com> References: <1265155168-28909-11-git-send-email-wellsk40@gmail.com> Message-ID: <20100203101802.GD22275@n2100.arm.linux.org.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, Feb 02, 2010 at 03:59:22PM -0800, wellsk40 at gmail.com wrote: > diff --git a/arch/arm/mach-lpc32xx/include/mach/platform.h b/arch/arm/mach-lpc32xx/include/mach/platform.h > index adc932b..93541f4 100644 > --- a/arch/arm/mach-lpc32xx/include/mach/platform.h > +++ b/arch/arm/mach-lpc32xx/include/mach/platform.h > @@ -23,6 +23,8 @@ > #ifndef __ASM_ARCH_PLATFORM_H > #define __ASM_ARCH_PLATFORM_H > > +#define IOMEM(x, y) (__force void __iomem *)((x) + (y)) Hmm, I'd prefer if this wasn't soo different from the definitions found in other platforms - so maybe one day we can move this to the arch level. The best definition for it is: #define IOMEM(x) ((void __iomem *)(unsigned long)(x)) which I believe avoids sparse complaints without having to resort to __force.