> > Does it mean pcilkm doesn't work with small changes on
> > NetBSD-current and we need another approach?
> 
> It might, but what's the point when a better way of writing PCI devices
> LKMs already exists?  Matthias's way is much better.  pcilkm was not a
> lot more than a hack.

What is `Matthias's way'?

-- 
TAMURA Kent <kent_2006 at hauN.org> <kent at NetBSD.org>