We were effectively assuming virtual memory in os-process targets was
contiguous... ouch.  Somebody (me) got a little too excited about just
calling the base target!

The fix for os_process_write is a little unfortunate, though.
Basically, we assume that if we can resolve all the v2p mappings and
read a byte from each page, that the page is present.  So we actually do
all those checks before writing anything.  These checks of course are
racy if the VM is not paused, but right now it's the best we can do.
The only other solution would be to hack the hypervisor page protection
bits, which would be super-heavyweight.  And we're not going to do that
anytime soon!