On Tue, 18 Jul 2000, Leo Smiers wrote:

> That helped. See the attachment.

Right.  That at least tells me that the unhandled faults (SIGSEGVs from
the user's point of view) are caused by the system running out of VM (line
1176 onwards in sys/uvm/uvm_fault.c).  That this is happening too soon for
init to be able to start indicates that we need to build a rather smaller
kernel for <8Mb systems, which is a bit of a pain.  I'll add it to the top
of todo.list.

There's still the question of that address exception to deal with, since
that shouldn't generally happen.  If you can reproduce it, the output of
"trace" at the db> prompt there would be useful, since it happened in SVC
mode.  Sadly, the debugger's output doesn't get saved in the message
buffer, but if you could record the function names (and the value of
"scp", the save code pointer), that should be helpful.

-- 
Ben Harris                                                   <bjh21@netbsd.org>
Portmaster, NetBSD/arm26               <URL:http://www.netbsd.org/Ports/arm26/>