VD: Driver
This version of the VD: driver for VMS fixes some major design flaws
in the 1986 version and now appears to work reliably, at least where
I've tried it (a VMS V4.7 system with directly attached massbus
disks). It uses the same method as MSCP so should also be reliable
on other systems. I added V5 code (conditioned by defining VMS$V5)
but can't test it. Tests here have been for one unit.
  Note that the one thing that could cause trouble would be if your
interrupt stack isn't big enough; during I/O postprocessing the
driver uses some interrupt stack. If you want to be extra safe, try
bumping an extra page into interrupt stack.