As a followup, I've found that this panic doesn't happen when I specify
a block size to restore.  ( i.e. 'restore ivbf /dev/nrst0 1000' on the amiga 
or 'restore ivbf amiga:/dev/nrst0 1000' on the sun).  The panic DOES
happen when I try to run restore without specifying a block size from the
amiga itself -- this isn't a problem caused by my sparc, it's in netbsd
itself.
So I can get by without causing panics, but having your computer crash when 
someone tries to use restore's default options is still not a good thing.   :-)

"'Most Excellent' Eric Mehlhaff" <mehlhaff> recently wrote:
>
>I have a sparc 5 running sunos4.1.3 which I'm trying to use rdump/rrestore
>to a HP35480 tape drive on my amiga3000 over ethernet. The dump part seems to 
>work fine, but I can't use remote restores from the sparc, because it always 
>causes a kernal panic on the amiga.  I'm basically just doing a 
>'restore ivf amiga:/dev/nrst0' from the sparc, and the amiga just goes down.
>
>The amiga's console message is:
>panic: vmapbuf: null page frame
>
>Which then drops me into the kernel debugger. A trace shows basically:
>
>vmapbuf()
>physio()
>rawread()
>spec_read()
>ufsspec_read()
>vn_read()
>read()
>syscall()
>trap0()
>
>( There's lots of numbers and such that I've omitted from the trace, but I can
> 
>supply them if they'd really be useful).
>
>The tape drive appears to work fine for local dumps and restores on the 
>amiga, with the exception  that 'mt -f /dev/nrst0 status' always reports: 
>unknown tape drive type (7)
>
>This is with a reasonably current kernel, compiled from last week's
>tarballs.  It also happened on on my older kernel compiled  from
>mid-February tarballs.

 Eric Mehlhaff, mehlhaff@crl.com