On Tue, 04 May 2004 23:43:26 +0100
Richard Earnshaw <rearnsha@buzzard.freeserve.co.uk> wrote:

> 
> > On Tue, 04 May 2004 23:14:00 +0100
> > Richard Earnshaw <rearnsha@buzzard.freeserve.co.uk> wrote:
> > 
> > > Chris,
> > > 
> > > There still seem to be numerous problems on the Cats with 2.0
> > > beta.
> > > 
> > > - Booting a non-standard kernel (eg at the firmware prompt typing
> > > boot
> > > 
> > > wd0:/testnetbsd.aout) seems to lead to the console being dead once
> > > 
> > > bootstrap is complete (you can type in the name of the root file
> > > system and init program, but you can't login afterwards).
> > 
> > How big is the kernel? if it's over 4 (or 5MB) it will start to boot
> > and then stop somewhere.
> 
> Nothing like that (about 2.5MB).  The kernel is fully functional (you
> can remote login), but the console keyboard is dead.

Ah I've seen this one, it only happens if you have to type in the root
etc.  I think it's a bug somewhere in the keyboard code.  It doesn't
seem to like switching from polling to interrupt mode.

One thing I know has cured it is when the X server attempts to take over
and fails (cos the X server doesn't work, and I can't fix it any more,
must send out email on that)

> > 
> > what version of cyclone is this?
> > 
> > > - Kernels built with LOCKDEBUG will panic very quickly once
> > > booted.
> > 
> > Where do they panic?
> 
> In the lock assertion code.  I'll have to try this one again, but I
> think it was "locking against myself".

I'll take a look if I get time.

Chris