[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Mint Addition



>BERNARD Sebastien writes:
> > Would it be interesting to include some of the driver into the mint code ?
> > I am proposing such a thing because I think that some features proposed
> > as extra should become standard to the kernel and use it as if.
> > [...]
> >   My proposal is to start a discussion about the extension :
> >     - should the extensions be kept extern to the kernel
> >       ou should they be integrated to the kernel and which ones ?
>I don't think that is good idea --- many MiNT-users don't need the
>(net-)extensions and the kernel would be bigger as needed.
>
>I prefer the opposite way --- remove more things from the kernel to
>drivers, the 'fasttext'-device for example (I never use it and it is
>not present in my kernel).  That makes the kernel smaller and you
>don't need to keep different versions of the kernel.

To do this completely we'll need functions to hook into the scheduler, be
able to get the kernel to vector to a device driver upon an interrupt from a
peripheral etc. and do this efficently.

We are obviously not there yet as we have to use kludgey solutions such as
the sld daemon for MiNTnet. The functionality of the sld daemon and other
such daemons should be able to be set up seemlessly using kernel callbacks
into device drivers etc. The kernel scheduler itself should be at a lower
priority than devices so that interrupts are handled as quickly as possible.

>Jens Krinke
>
>____________________________________
>Jens Krinke (krinke@ips.cs.tu-bs.de)

Steve

-- 
---------------------------------------------------------------------------
Computer Systems Administrator, Dept. of Earth Sciences, Oxford University.
E-Mail: steve@uk.ac.ox.earth (JANET) steve@earth.ox.ac.uk (Internet).
Tel:- Oxford (0865) 282110 (UK) or +44 865 282110 (International).