Le Sat, 9 Nov 2002 21:23:14 +0000
Rasputin a écrit :
> * vodka@intoxicate.us <vodka@intoxicate.us> [1146 23:46]:
> > > At 09:58 PM 11/8/2002 +0100, Tld wrote:
> > 
> > > I think its more involved than that. NetBSD-core may have negotiate
> > > with
> > >  them. The porting is usually not the problem but license. The
> > >  answer
> > > from  Nvidia would depend upon if they think there is business.
> > 
> > Hmm, I wonder if they really make these drivers to increase their
> > business.  I can understand why they have Linux drivers, since a good
[...]
> I think FreeBSD have the FreeBSD driver Initiative nagging NVidia for
> years to thank as much as anything. 
> 
> Personally, I don't see why NVidia would have a problem with NetBSD
> using the code, since they've already given it away once, effectively.
> In theory it would open up more markets (I was thinking of PowerPC in
> particular).
> 
> This all assumes it's not too FreeBSD specific, of course (i.e. doesn't
> need DRI).

I spent the day digging through NVidia, FreeBSD and NetBSD sources, and as
long as I'm not a too dumb hacker, the driver is not too FreeBSD specific.

Of course, there is a lot of issues, but they mainly involve memory
management. One of them is the use of FreeBSD's contigmalloc, which
doesn't seem to have an equivalent in NetBSD. 

I still haven't checked if resource allocation is actually possible within
a LKM, but I've seen no reason against it, yet.

I'll post on tech-kern when I'm done gathering all my questions about the
technical aspects of the porting.

Last but not least, one must check if nvidia_drv.o is compatible with
NetBSD's XFree86.

cube

-- 
That's how I did it, Anton. I never saved anything
for the swim back.
Vincent/Jerome, in Gattaca.