On Thu, 23 Jul 1998, Neil A. Carson wrote:

[reason for UVM-related panic]

> anything obvious, but I'd think that, if it was obvious, it'd have
> happened to others too. How many other people are reporting problems
> with UVM, and on what machines?

This might be a red-herring, and totally unrelated but... I have three
machines running -current (1 x mvme68k, 2 x Shark). The mvme68k is pretty
solid (especially since the descriptor leak in sys/kern/subr_extent.c was
plugged last week). However, the two Sharks are unstable. One is fairly
unloaded; just acting as a NAT modem gateway for my home network. The
other is used fairly heavily; X, Mozilla etc. The first Shark reboots
itself after about a week of uptime, the second Shark 'locks-up' readily
during heavy paging.

Unfortunately, the first Shark has no monitor so I can't see the reason
for the reboot (panic?) and the second just locks up. In both cases, the
reboot wipes RAM so the message buffer doesn't give anything away. :( I'll
attempt to set up a serial console on the second Shark soon, when I get
hold of a PS/2 mouse.

Like I said, it might be a red-herring; both Sharks might be dodgy!

Cheers, Steve