> In article <199811111858.TAA03864@gloin.fi.muni.cz>,
> Jan Kasprzak <kas@informatics.muni.cz> wrote:
> >
> > Can anybody explain this to me? Is (x)-PAGE_OFFSET at any chance
> >slower or less efficient than (x)&~PAGE_OFFSET?
>
> No.
>
> But the one using logical operations is safer in the presense of old
> drivers that don't do the right thing. There are some historical
> drivers that may mix physical and virtual addresses, and just masking
> off (or turning on) the high bits makes them work even though they
> shouldn't.
>
> For 2.3.x we definitely want to use the +-PAGE_OFFSET version.
And we probably will implement these ones differently:
#define readb(addr) (*(volatile unsigned char *) __io_virt(addr))
#define readw(addr) (*(volatile unsigned short *) __io_virt(addr))
#define readl(addr) (*(volatile unsigned int *) __io_virt(addr))
#define writeb(b,addr) (*(volatile unsigned char *) __io_virt(addr) = (b))
#define writew(b,addr) (*(volatile unsigned short *) __io_virt(addr) = (b))
#define writel(b,addr) (*(volatile unsigned int *) __io_virt(addr) = (b))
#define memset_io(a,b,c) memset(__io_virt(a),(b),(c))
#define memcpy_fromio(a,b,c) memcpy((a),__io_virt(b),(c))
#define memcpy_toio(a,b,c) memcpy(__io_virt(a),(b),(c))
The +-PAGE_OFFSET version breaks all the above functions when the address
is remapped using ioremap(). (PCI for example).
In practice, the current implementation of x86/io.h allows to deal with
high addresses and low addresses (and probably old drivers) by
transparently working for addresses that are remapped and those that are
not. But the cost is 1 useless logical operation for each MMIO access.
Regards,
Gerard.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/