On Sun, Jun 06, 1999 at 12:02:55AM +0200, Markus Kilbinger wrote:
>   (gdb) target kcore netbsd.5.core
>   panic: kernel fault
>   #0  0xf0033d7c in mi_switch ()
>   (gdb) bt
>   #0  0xf0033d7c in mi_switch ()
>   #1  0xf003379c in tsleep ()
>   #2  0xf00de648 in uvm_scheduler ()
>   #3  0xf002347c in main ()
> 
I used to get these on my IPC as well, though it's been some time I didn't
see them.
If I remember properly, it fails at the cpu_switch() call in mi_switch(),
gdb tells me p is a valid pointer.
As cpu_switch() is a assembly routine, I don't know if the fault really
happens at the function call, or happens in cpu_switch() but it didn't
leave a frame for the debugger.
One suggestion I made at that time was that the stack space was exhausted
and the fault occured while trying to write to the stack for the function
call, but I never got an anserw about this (I don't know how to confirm/infirm
this, I don't know the sparc assembly).

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--