On Mon, Dec 09, 2002 at 01:56:36PM +0100, Eduard Bloch wrote: > #include <hallo.h> > Bernhard R. Link wrote on Mon Dec 09, 2002 um 11:44:28AM: > > > This design decision could exclude the segment between desktops and > > embedded things, where the only ways out are are serial-cards (and > > maybe some other things, for which getting a initial console would be > > even harder). > > Maybe. And the solution is: fix the kernel to the right time. There is large > segment between desktops and embedded things that works fine with Devfs, > please do not try to scare us. > > > I'm not saying devfs should not used at all. But the system should > > support using kernels without devfs, too. > > The sense behind switching to Devfs is exactly this problem - getting rid of > painfull work of creating and maintaining the device files. Who is going to do > it? You. Feel free to do. I recommend switching to Devfs now and fixing its > bugs later. As soon as you require devfs, you drop support for platforms that don't have working 2.4 kernels. I have hopes of getting 2.4 working for mac68k one day, but I'd hate to lose installation support for at least four of our m68k buildds. People use these boxes, there just aren't a lot (okay I think about one guy) of kernel developers. I seem to recall some sparcs have this problem too. -- Stephen R. Marenka If life's not fun, you're not doing it right! <stephen@marenka.net>
Attachment:
pgpR4VdeChoYi.pgp
Description: PGP signature