On Fri, 5 May 2000, Horst von Brand wrote:
> Tigran Aivazian <tigran@veritas.com> said:
>
> [...]
>
> > Ok, I started to work on it but on the back of my mind there is still
> > one unanswered question - do we force nullfs to be mounted somewhere
> > (/nullfs ?) automatically on boot and make this little filesystem required
> > non-modular built-in thing or do we make it modular and make sure users
> > mount it if they need it (on the other hand *everybody* needs forced
> > umount, e.g. to shutdown their systems if processes don't die in peace).
>
> How about overloading chdir(2) so that chdir(NULL) just moves the processes
> cwd to the (or a) non-directory? I don't believe there are valid cases of
Stop it, please. It's just plain ugly.
> chdir(NULL). Another possibility could be to use chroot(2), which forbids
There are none and there is no valid case for overloading the basic interface.
> the use by non-root (good idea, IMHO) in the same call. The non-dir cage
> doesn't need to show up in the filesystem this way, thus killing the
> posibilities of mistakes and screams about strange stuff being
> automagically mounted. Makes sense too as it isn't really part of the
> hierarchy.
We don't have to mount it anywhere within the user's reach. I'ld like to
see Tigran's code before commenting on his design, but the kludgery above
is just too ugly - it reeks with Missed'em'V if not worse.
-
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 : Sun May 07 2000 - 21:00:17 EST