Subject: MMU Crashes.
To: None <port-sun3@NetBSD.ORG>
From: Wayne D. Hoxsie Jr. <firstname.lastname@example.org>
Date: 03/06/1998 09:34:30
I finally got netbsd working on my sun3; however, the machine periodically
(sometimes just while unattended) will trap with an MMU error of some sort
(don't have a dump copy handy) mentioning virtual memory or the like.
I'm wondering if I have my swap space set up properly (and the main
partition for that matter), the reason being, I could still boot from the
installation filesystem leftover on the swap partition even several days
after the installation. This seems to indicate that it is not using the
swap partition as it surely would have been overwritten. Also, I've done
"continue" from the "db >" prompt after a trap and it dumps all over my root
partition and scrambles the boot kernel necessitating a complete reinstall.
My partitioning is set up with a 32 meg swap from 1st cyl to (+32 megs),
then the rest is my main partition from block (+32megs + next cyl boundary)
to the end. (I don't have the exact parameters handy)
Should the swap be at the begining or end of the physical disk or does it
matter? Is there any other place the kernel could be getting partition
information and thinking its different than what edlabel says?