Re: [PATCH 7/8] ARM: OMAP: Remove plat/io.h by moving it into mach/io.h

[Tony Lindgren <tony@atomide.com>]

* Rob Herring <robherring2@gmail.com> [120221 17:53]:
> On 02/21/2012 05:40 PM, Tony Lindgren wrote:
> > This is needed so we can eventually remove mach/io.h
> > for omap2+.
...

> > -#include <plat/io.h>
> > +#ifndef __ASM_ARM_ARCH_IO_H
> > +#define __ASM_ARM_ARCH_IO_H
> > +
> > +#include <mach/hardware.h>
> 
> This shouldn't be needed.

Good catch, will update things for that.
 
> > +
> > +#define IO_SPACE_LIMIT 0xffffffff
> > +
> > +/*
> > + * We don't actually have real ISA nor PCI buses, but there is so many
> > + * drivers out there that might just work if we fake them...
> > + */
> > +#define __io(a)		__typesafe_io(a)
> > +#define __mem_pci(a)	(a)
> > +
> > +/*
> > + * ----------------------------------------------------------------------------
> > + * I/O mapping
> > + * ----------------------------------------------------------------------------
> > + */
> > +
> > +#ifdef __ASSEMBLER__
> > +#define IOMEM(x)		(x)
> > +#else
> > +#define IOMEM(x)		((void __force __iomem *)(x))
> > +#endif
> > +
> 
> This can't be moved to iomap.h? I did that for other platforms.

Well I was thinking we should define these somewhere for everybody,
so I was trying to avoid adding yet another copy.. But I can move
it and then we can clean it up later to avoid more dependencies here.
 
> > +/*
> > + * ----------------------------------------------------------------------------
> > + * Omap specific register access
> > + * ----------------------------------------------------------------------------
> > + */
> > +
> > +#ifndef __ASSEMBLER__
> > +
> > +/*
> > + * NOTE: Please use ioremap + __raw_read/write where possible instead of these
> > + */
> > +
> > +extern u8 omap_readb(u32 pa);
> > +extern u16 omap_readw(u32 pa);
> > +extern u32 omap_readl(u32 pa);
> > +extern void omap_writeb(u8 v, u32 pa);
> > +extern void omap_writew(u16 v, u32 pa);
> > +extern void omap_writel(u32 v, u32 pa);
> 
> These can't be moved to common.h like omap2?

Not for omap1 currently because a bunch of legacy drivers are
using them :(

Anyways, that should be OK for now as omap1 requires v5 flags
and it currently does not make sense to try to compile into a
common zImage. At least the USB driver already needs to be
cleaned up for that, so eventually omap1 io.h can disappear
too.
 
> Same comments on omap2 io.h.

Thanks, will post an updated patch.

Regards,

Tony