> > As a side point, is it worth implementing a much more complex
> > configuration manager as a user-side daemon, and have a very simple
> > configuration manager in the kernel itself?
>
> This seems the best approach.
>
> As for the rest of the issues, they _are_ quite tricky. For example with
> my NodeRunner, it _can_ be reconfigured, but probably not without
There are quite a few of these types of devices around. The general idea
is that the 'run-time configuration daemon' should be able to make
decisions about whether it is worth rebooting because it is getting in the
way of something vital, or whether it can be worked around.
> rebooting, and as it's an EEPROM, it probably doesn't have that many
> available write cycles. So it would be wonderful to be able to plug in a
> module that lets the configuration manager _know_ it can be reconfigured,
> and do so it if I explicitly authorize it. But doing it without permission
> would be a no-no, and ideally there should be a weighting to suggest that
> changing that device is less favorable then changing other devices.
>
> I'm wondering if a Prolog engine (or something along those lines) would be
> needed for the full-scale configuration manager.
I don't know enough about Prolog to say much about this, but I suspect
that the only way of doing it is by a 'brute-force' approach - which is
the reason I would prefer not having it in the kernel in the first place.
Another question is, do we listen to what the PnP BIOS has already set up
for us if it's there, or do we do things our own way?
-- Tom Lees <tom@lpsg.demon.co.uk> http://www.lpsg.demon.co.uk/ PGP ID 87D4D065, fingerprint 2A 66 86 9D 02 4D A6 1E B8 A2 17 9D 4F 9B 89 D6 finger tom@master.debian.org for full public key (also available on keyservers)