Hiya Stephen,

On Sun, Apr 07, 2002 at 05:45:24PM +0100, Stephen Mollett wrote:
> I'm sure I must be overlooking something really simple and obvious, or just 
> doing something stoopid :o) Having not managed to get a kernel to boot 
> successfully on an NC (see my previous message, "Bootloader problems on an 
> NC" - I've now tried several NCs to rule out faulty hardware) I decided to 
> try installing NetBSD on my RiscPC 600, partly to simply get myself a NetBSD 
> system to play with and partly so that I could try compiling a new kernel for 
> the NC from the latest source.

hmm.... i'll dig into the history to find that message... i must have 
missed it :-/

> Using BtNetBSD 0.99a, set up according to the docs, together with the 
> installation kernel, I tried to boot into NetBSD to begin installation. The 
> loader found the memory OK (16M+16M+2M), did the various memory-mapping and 
> relocation stages then did the 5-second countdown. Then the system locked 
> solid when the loader said "start!"

ah... latest sources you said i.e. -current?

In my 1.6 roundup checks i've encountered the same problem... its fixed now 
... there was a processor configuration function going astray that killed 
the processor.

You should update the sources to today and then rebuild the kernels... they 
ought to work now ... please kill off the old compile directory first 
please and reconfig ...

> I've also tried various other kernels for the RiscPC, and a generic one, with 
> exactly the same result. The same thing happens if I boot to the supervisor 
> prompt then run "!BtNetBSD.BtNetBSD $.netbsd" directly. (As an experiment, 
> I've tried one of the RiscPC kernels on an NC just to see what happened; 
> exactly the same result. I don't know whether one would expect to get any 
> further than this, though.)

The difference between the two is not that big and one day they even ought 
to be able to boot one kernel.... same problem as above...

I'm currently building a complete snapshot to see if all is going right and 
everything works fine.

Cheers,
Reinoud