> Date: Tue, 19 Nov 1996 17:27:36 -0600 (CST)
> From: Henry W Miller <mill0440@gold.tc.umn.edu>

> I'm netbooting a sun 3/60 from a linux box, and the netbsd-gen kernel is 
> dieing at a repeatable point in the boot process.  I have a cg4 
> framebuffer in this machine, but there is no monitor or keybaord 
> attacthed, I'm using port a as my console.
> 
> em0 at mainbus0
> cgfour0 at obmem0 addr 0xff200000 (115
> 
> At this point everything stops.  pressing enter brings up a db> prompt.  
> I assume I'm already in teh bugger.  On a cold start the last few 0's of 
> the memory address will be garbage, if I press c and the db prompt it 
> reboots, and gets to the (  then there is 3-4 charters of random garbage.

That is strange...  Does the "db> trace" command work?
If so, you could send it to me (privately) and maybe I
could make some guesses about why this happens...

> I should also mention that if another machine trys to ping this box 
> exactly one ping is responded to before it dies.  If two machines are 
> painging it (one ping/second) generaly only one gets a response, rarely 
> both get a single reply.  
> 
> If I use the netbsd-rd kernel I get a ssh> prompt, and it seems to wrok, 
> but this isn't useful for netbooting.

Are you sure the hardware is OK?  (Passes memory diagnostics?)
The ramdisk kernel might use less of the memory...

> Any ideas?  I'm willing to run a different kernel, so long as it is 
> stable enough to run a xserver.

Maybe you could see if SunOS works correctly on the hardware...

If it turns out to be a NetBSD bug I'd like to know.

Thanks,
Gordon