On Wed, 26 Jul 2000, Michael T. Gilmore wrote:
> Mounting (or replacing) on an current mount path, busy or nor, wil really
> lead to bizzare and damaging senerios later.
"Doctor, it hurts when I do it"
> You may not see the "why" now, but when your process mounts a volume, and
> takes time between acess (busy), if it is replaced, something will go boom!
Yes? And if it cares to do umount first? _If_ you have processes running
around and fooling with mounts - keep the mounted stuff busy, you have no
other warranty. On any UNIX. Or, better yet, don't have such processes.
> RULE: once a path is mounted upon, no mounts are allowed on any of the path
> (directory) components.
Too bad, a lot of Unices do not follow that rule. Check yourself. Linux
(since _very_ long, if not from the day 1) too. Heck, IIRC v7 allows to
mount on /foo/bar/baz and then mount on /foo.
> If you insist on adding this "feature", please make it a system tunable (config)
> option. Some of us really do not want it.
Lack of that does not add you _any_ safety. Again, if fs is not busy - it
_can_ be replaced. Takes two system calls instead of one - BFD, they
require exactly the same privileges. It may be a problem, but it's _not_ a
security problem. It does not add any new privileges - same effect could
always be achieved by combination of umount() and mount().
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Jul 31 2000 - 21:00:22 EST