On Fri, 9 May 97 00:09:50,
is@beverly.rhein.de (Ignatios Souvatzis) wrote about Re: Starting of CV6=
4/3D driver:
> Hi Sebastian,
>=20
> > >  Everybody needs a start. Writing a device driver is not any harde=
r
> > > than a decent numeric or stochastic method. It doesnt require
> > > lot of brain - just some "read the source, Luke" (other device
> > > drivers) and infos about the chip set of the graphic board (which =
are,
> > > in the case of Phase5, not really been made available...)
> >=20
> > What should I do? Recource/reassemble CVision3D monitor? Ummm...
>=20
> Oh... no.
>=20
> First of all, read source code of other drivers. Read "The design and
> implementation of the 4.4BSD OS" book for background, although in NetB=
SD
> there are some changes. Read the section 9 manpages (and yes, if you a=
re
> prepared to rebuild your system sometimes, upgrading to some current
> version might be a good think, if you are going to write drivers).

Thank you very much.
=20
> In case youre unsure about this: you can use NetBSD on your ECS/AGA ev=
en
> if your CV3D isn't supported, for the driver development.

Im using NetBSD very often on my A4000. I know it ;)

> Before you actually start, ask Phase5 about board specifications: whic=
h
> chip is at what address offset relative to the board base; where the
> framebuffer memory is, which documents to read about the chips.

I think P5 don't provide me with docs. I have heard that author of CV64
driver for NetBSD and Linux had to resource CVision monitor. Maybe they
give me a chance? :)

> Normally there shouldnt be the need to reverse engineer stuff.

On S3's WWW server there're only brief reviews in PDF or mikrozoft's Wor=
d
format. I couldn't find any technical reference(s) for the ViRGE chip.
=20
--=20
Sebastian Krawczuk                                stager@dedal.man.szcze=
cin.pl
                 GodLess Zine by K.Jablonski at http://www.szczecin.pl/~=
stager