In article <MeV8_XO00WBKQ34Uo8@andrew.cmu.edu> fl0p+@andrew.cmu.edu (Frank T Lofaro) writes:
quoted text
>    Has this been fixed? Is it just a problem with my setup? Is it my
>kernel, my mount, my fs (ext), or what? I'm using the mount I got from
>the 0.96 rootdisk. It does say (nosuid) when I check the mount options
>just by checking mount. Would it say that it mounted okay and nosuid,
>store that in /etc/mtab, even if it didn't actually implement the
>nosuid??? (the right behavior would be to not mount, or at least
>complain, or at the very least not say it is nosuid when it isn't).

This is a problem with the mount on the 0.96 root disk - it was
written before the nosuid, nodev, etc filesystem options were added to
the kernel, and ended up using both different numbers and interface
than the was implemented in the kernel.  So, it is not a problem once
you have a 'correct' mount.

The mount on the 0.97 root disk should suit your purposes.  The
problems that version has is with passing options to MS-DOS
filesystems, but it works fine for extended and Minix partitions.  I
will release a 0.97.1 root disk in a week or so that will have a fixed
mount, and it will also correct some other problems.

quoted text
>    The mount program seems a little hokey, it doesn't seem to have very
>good error checking, maybe that is part of the problem...

Hmm, I wonder why you call it hokey - it is a pretty straight port of
the BSD mount, minus the NFS stuff, for the most part.  The problem is
in the way it was originally assumed mount(2) would be implemented and
the way it actually was.

quoted text
>    By the way, what is the right behaviour for a (correctly working)
>nosuid option when a user tries to run a suid progam off a nosuid
>partition? Say permission denied, or let the program run without
>setuid/setgid permissions? Does POSIX have anything to say on this?

The response I got from 'su' on a nosuid filesystem was 'su: operation
not permitted'.  I think what happens is the program just gets run
with 'normal' permissions (none of the setuid is done), and it will
only break if it really needs the extra permission.  That's just
speculation (a swag, as it were), however, and I don't know what POSIX
has to say about it.
-- 
                                    +    Jim Winstead Jr. (CSci '95)
                                    |            Harvey Mudd College
                                    | jwinstea@jarthur.Claremont.EDU
                                    + This is all my words.  Honest!