Re: introducing devmajor_t/devminor_t

[David Brownlee <abs%NetBSD.org@localhost>]

On Thu, 15 Jan 2009, Matthias Drochner wrote:

> Here are patches which make major()/minor() return a
> 32-bit value again, which avoids 64-bit arithmetics
> where major and minor numbers are used as array
> indices. This would also make sure that device
> unit numbers in drivers are still plain integers,
> so no PRI64s are needed to print out device instances.
>
> I've introduced new type names for this. This doesn't
> make a difference technically but serves as a kind
> of comment. (Some OSes have this too, or major_t/minor_t,
> so the idea is not that exotic.)
>
> What do you think?

        It seems pretty sane - and cleans things up nicely.
        Out of curiosity what type names do other systems use?

--
                David/absolute       -- www.NetBSD.org: No hype required --