> One way is to put the export flags, etc in every i-node, but that's a lot
> of work and you can have fun propagating them down the directory tree.

Oops, I realized after posting that this was very poorly worded. What I
meant was something like:
- If there are export flags on every i-node, then the only problem is how
  you set them for the file with multiple hard links (what I implied by
  the "have fun propagating" nonsense above).
  --> You can maintain security by setting them to the least permissive
      case for the various hard links (ie. ro for the case in the previous
      post).
  This has the weird effect that there would be a few files in a subtree
  that would report EROFS when a write is attempted, even though the file
  subtree was mounted read/write.

Sorry about the obscure statement. I'm in "teaching mode" these days and
like to give students part of the answer, so they have to thing about the
rest:-) rick