What to do after a kernel panic

[John Klos <john%sixgirls.org@localhost>]

Hello,

I have a machine that will not be running anything that's very critical,
but I will definitely not be able to get physical access to it once it's
installed (it's getting installed in a very weird location - I'll send
more info on this a little later).

It's an Amiga 3000, 68030, 16 megs of ram. For some reason, I got kernel
panic while doing a make build in /usr/src (I have just 1.5 and the
regular 1.5 source tree; I was doing the build to test the machine). The
kernel is pretty much a GENERIC kernel that had maxusers set to 16 and all
of the hardware that I do not have commented out. (I have since changed
maxusers to 8).

While I do care about why I got a kernel panic, I am much more interested
in knowing this: is there any way to build a kernel that does something
other than stopping when a panic happens? Even if there is no way to get
the panic message (perhaps it can go to the serial port), I'd love to have
the machine reboot itself rather than stop.

Unfortunately, remote reboot is not an option. Getting access to the
serial port may be an option, but it'd be a lot of extra work. So the best
thing would be if the panic was sent to the serial port and the computer
rebooted automatically.

Any ideas?

Thanks,
John Klos
-- 
Those who do not understand Unix are condemned to reinvent it, poorly.
                -- Henry Spencer