On Fri, 2006-03-17 at 11:21 +0000, Bahadir Balban wrote:
> Hi there,
>
> To access device registers, which one is best practice in terms of
> coding style, and is any of them totally unacceptable for a driver to
> be merged to the kernel?
>
> 1) Is this ok?
> static inline unsigned int device_read_reg(int offset)
> {
> return readl(device_global_struct->base + offset);
> }
this sounds bad. Why? because this can only ever work with 1 piece of
hardware. Can you 200% guarantee there never ever will be systems with 2
of your hardware?
Also such wrappers are generally a bad idea, they just add no value
other than obfuscating your code and making it harder to read.
> 2) I've seen this in some kernel code, note the &:
> readl(&device_global_struct->registeroffset);
> writel(&device_global_struct->registeroffset, data);
>
> 3) Also possible:
> readl(device_global_struct->registeroffset);
> writel(device_global_struct->registeroffset, data);
these are 2 different uses
in the first case (2) device_global_struct is a *fake* pointer which
"points" to the ioremap cookie. The "& ... -> .... " construct is a
fancy way to write "+" to the compiler basically.
in the second case (3) device_global_struct is a real structure, in
which you stored the cookie itself.
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
