Re: [MiNT] This fight

[Petr Stehlik <joy@sophics.cz>]

On Thu, 2003-12-11 at 13:18, Odd Skancke wrote:
> if this can apply to all NF, but.. Lets take Shutdown(), as this is a
> standard, generic kernel function. The kernel should then use the external
> modules to actually perform the hardware part of the shutdown. 

right.

> the impression that NatFeats should NOT be used by user applicaitons (only
> the OS). Why then do we need to make MiNTlib aware/use NatFeat?

For example there is a MiNTlib function that returns the name of machine
it is running on (sorry I can't recall it ATM) - it returns "Falcon" or
"TT" or things like that. I think it would be good if it returned
"ARAnyM" for ARAnyM, "STonX" for STonX, "hatari" for hatari etc. 

Now I suppose the MiNTlib would wrap the FreeMiNT's proposed NF syscall
in a neat function call and would use this call in the routine that
finds out the machine name. But in TOS emulation mode (when MiNTlib
linked program would be running on TOS/MagiC) it should fall back to
__NF cookie.

Another option is to write TSR for TOS that would provide emulation of
the MiNT syscall in TOS.

> > But if you agree with my view (and agree with making FreeMiNT and
> > MiNTlib aware of NF) then the __NF cookie could simply be removed from
> > the CookieJar upon MiNT boot so you'd get a clean system with no
> > pointers.
> 
>  I think I need to know more about NatFeats. I thought it could be
> regardes as "new/different hardware". In which case support for it should
> go in XDD's, and only there. Not in MiNTlib or anywhere else.

There were proposals by other guys to use NF for multimedia processing,
for example (JPEG as the simplest example).

>  Where can I read about what NatFeats really is?

I posted a direct link to NF Doc at 11:45 today (subj. "Shutdown()
discussion").

Petr