"Chris Swiedler" <chris.swiedler@rememberit.com>:
> > > I think it would be most intuitive for users if we can manage to have
> the
> > > floppy automounted when it is inserted and autounmounted when removed.
> Dirty
> > > data needs to be flushed expediently for this to work well.
> >
> > And what user owns the floppy ? Everybody who is logged in?
> > Can anybody write to it? Can everybody write to it.
> >
> > It the user who put the floppy in the drive stores some sensitive
> > data (X.509 certificates come to mind) who can read it?.
>
> These are more reasons why I suggested a single-user mode as the answer. If
> we assume that the only user is sitting at the computer, then we can
> restrict priveledges to that user.
And how do you determine that. The user may have logged into X, still at
the console, but using pty for command line.
> To expand the idea a little: trying to rewrite Linux as a single-user OS is
> obviously extreme. What if instead we had special recognition for the user
> at the console? I.e., if the user has logged in via the terminal itself (not
> a serial connection, telnet session, or anything else) then we can assume
> that floppies which are inserted can be automounted, and privledges
> restricted to that user alone. It certainly makes sense--how is someone
> other than the console user going to insert a floppy? If there is no console
> user logged in, the floppy must be mounted manually as usual. If the floppy
> is really intended for a remote user, then the console user (if any) must
> manually unmount the disk so that the remote user can mount it. If the disk
> is removed prematurely, we know exactly who to notify.
>
> How difficult would it be to put code in the kernel like
>
> if (current->uid == console->uid)
> {
> /* special case for console user */
> ...
> }
console->uid is always 0 if XDM is being used.
-------------------------------------------------------------------------
Jesse I Pollard, II
Email: pollard@navo.hpc.mil
Any opinions expressed are solely my own.
-
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 : Fri Jun 23 2000 - 21:00:20 EST