Re: [MiNT] New kernel features

[Guido Flohr <gufl0000@stud.uni-sb.de>]

Hi,

On Sun, Dec 05, 1999 at 02:33:42PM +0100, Jo Even Skarstein wrote:
> On Sun, 5 Dec 1999 01:28:44 +0100, Guido Flohr wrote:
> 
> > If there is one thing I have learned from all that, it's this:  I think it
> > is about time to really split up the kernel development into two different
> > branches.  One, with minimal functionality, no features other than those
> > that are strictly required to do multitasking and one which is oriented
> > towards new developments in other operating systems.
> 
> Is it necessary to split the *development*? Frank has done a great job on
> reorganizing the kernel, and one side-effect of such work is usually that
> it's easier to separate functionality in modules. If new features can be
> ifdef'ed like your kern-fs, then it should be trivial to release slimmed-down
> kernels. Also, the mere possibility of doing this indicates good design and
> easier maintenance.

That was what I had in mind.  The term "split up" was not correct, sorry.

> Btw. this also means that the "slim" kernel-binaries should be distributed
> along with the ordinary kernels, I really don't like the idea that the users
> should have to recompile the kernel themselves, after all, anyone who needs
> a slimmed-down kernel probably don't have the resources to compile it. It
> would actually be better to just distribute "slim" kernels and require the
> "power-users" to recompile...

That's exactly what I meant with "natural selection".  The MiNT-Lite
kernel will be sufficient to compile the full-featured kernel.

Yes, loadable modules would be the best solution.  But seen from a
realistic point of view, the required changes are so radical that this
could only be done by the maintainer himself (at least he has to be
involved a whole lot).  I know that loadable modules are on Frank's TODO
list but probably not listed under items 1-3.  That has to be respected.

Ciao

Guido
-- 
http://stud.uni-sb.de/~gufl0000/
mailto:guido@atari.org