> After a fresh install of NetBSD-1.3 I found a real unexpected behaviour:
>
> I'm using an AS200 with 64MB ram , S3-Trio VGA console. In order to build
> a specific kernel, I wanted to print the file where all options for alpha
> are listed. printcap was not set up properly so I did a simple
>  'cat ALPHA > /dev/lpt0' ... whoa ! after 1 second the system was brought
> to SRM console prompt.
> The kernel? wrote : halted CPU 0
>                     halt code = 2
>                     kernel stack not valid halt
>                     PC=18a98
> any idea what went wrong?

Interesting!!!  This this repeatable?

Problems have been reported like this on the i386 (dating back ... a
while now) occasionally, but no solution has ever been found.

I assume this is with the standard lpt configuration (interrupt
driven).  You might see if the problem occurs if you use the driver's
polling mechanism (remove the 'irq N' from your kernel config line for
the lpt0 device).


> is the kernel debugging possible with the supplied generic kernel ?

Maybe, but a kernel debugger isn't going to help you much here anyway.
The fact that you're dropped into the firmware with 'kernel stack not
valid halt' indicates that the system was in a state where the normal
fault-handling (and debugger-invoking 8-) mechanisms wouldn't work.

examining the registers with the SRM 'examine' command might give some
useful information on where the crash happened, but i'd not bet too
much on it.  8-)


Unfortunately, i've never had a device that I could attach to any of
my machine's parallel ports to see if i could experience the problem
myself.



cgd