Kai Henningsen wrote:
> %d for public root-changeable int
> %?d if root can only read
> %0d if mortals can't read
> %?0d for both
>
and then %#$!!¤$$ for something deep rude.
Why can't struct net_device and other
few hundred to thousand different
non-performace sensitive structures
in kernel just have:
Simple virutal method table
aka metadata table pointer
describing value-fields,
methods, substructures,
access methods, permissions
in them, for example ?
Like it is in object-oriented language
C level implementations.
And /proc, khttpd, sysctl, ioctl, devfs,
whatever-comes-next-interface
just go pointer-walking trough
those hierarcial structures.
And single daemon making both data
and metadata persistent.
Plus, a lot of data property tables could be
made auto(un)loadable on demand.
Essentially all of those currently existing
interfaces are just interfaces to objects.
manipulating them one or another way.
Why to create just another messy interface
for which the code must be: created, changed
converted, transported, compiled, loaded,
maintained and thrown out ?
PLeez, do it the way, that if I want to use it
in device driver, then I could be real sure
that I never need to rewrite my code.
And, next innovative people playing with
every kind of interfaces, should not be waiting
on me and hundreds of other device driver
writers to do some consistent and covering-it-all
great stuff.
elmer.
-
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/
This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:27 EST