Bernd Sieker wrote:

> On 03.07.02, 14:45:04, Gregory McGarry wrote:
> > > The resulting kernel finds the framebuffer and makes its kernel output
> > > there, but when trying to boot, it complains:
> > > 
> > >   root file system type: nfs
> > >   init: copying out path '/sbin/init' 11
> > >   Jul  6 15:57:28 init: can't get /dev/console for controlling terminal:
> > >     Inappropriate ioctl for device
> > > 
> > > And on and on, until I reset.
> > 
> > You need to populate /dev on the nfs server.
> 
> Yes, I thought so, too. But I did that, ("./MAKEDEV all" should
> suffice, should it not?); and I can boot fine to stable-running
> multiuser when using remote console, but not on the framebuffer
> (console hardwired to code 133).
> 
> I have /dev/console, and also /dev/grf0 and /dev/ttye0 (which I guess
> is the text console device, it is make by "MAKEDEV ite*", ite being
> the kernel config entry for the "internal terminal emulator" on the
> framebuffer.)

OK.  Thanks for the report.  The problem was caused in the EPASSTHROUGH
jumbo patch.  Here is the fix to -current.  I have requested a pull-up
to 1.6.

> Module Name:	syssrc
> Committed By:	gmcgarry
> Date:		Fri Jul  5 23:43:29 UTC 2002
> 
> Modified Files:
> 	syssrc/sys/arch/hp300/dev: ite.c
> 
> Log Message:
> Invert test of EPASSTHROUGH.  Problem reported by
> Bernd Sieker <bsieker@freenet.de>.

	-- Gregory McGarry <g.mcgarry@ieee.org>