Re: Loopback (dir->dir) mounting (NOT -o loop)

From: Alexander Viro (viro@math.psu.edu)
Date: Wed Jun 07 2000 - 02:37:32 EST


On 7 Jun 2000, Nix wrote:

> Unfortunately, there are frequently administrative reasons to say,
> variously,
>
> `everyone unmount this right now, we need to take this disk offline'
> (chance to everyone's namespace)

Yes, but for that you need a working revoke() anyway and with that it
becomes trivial...

> `mount this so that all my daemons can see it' (change to namespace
> unrelated to yours)
>
> `everyone mount this right now, that disk is replaced and /opt/glurk is
> back where it should be'.
>
> I don't think automounters suffice for all of this --- and automounters
> are arguably overkill for this sort of thing, anyway. Is there a way to
> say that, if some processes' namespaces are clone/separated from all the
> others?

In principle, yes. At least if you give root a way to access others' mount
IDs (sorry, long story). BTW, automounter can be done seriously lighter -
search for "mount-traps" on fsdevel.

-
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 : Wed Jun 07 2000 - 21:00:27 EST