Re: NFS question
H.J. Lu (hjl@lucon.org)
Thu, 1 Oct 1998 07:45:23 -0700 (PDT)
>
> > > I think you may have misdiagnosed the problem Alex is reporting. What may be
> > > happening is that the file is executed on the client machine (as its permissions
> > > allow), and is then cached on the client side by the page cache. A subsequent
> > > attempt to read the file succeeds because of the cacheing.
> > >
> > > If this is what's happening, the fix needs to be made on the client side.
> > >
> >
> > No. It is a kernel nfsd bug. Otherwise, we are saying our kernel nfsd
> > only works with Linux client if you export a fs with root_squash.
> >
>
> I've tried your last patch. Now the 711 files could be executed, but after
> execution they could be read. To help understand where the problem is,
> I give another example:
> Assume I have 644 root's file exported to outside. From client machine
> I do cat that file (it runs unfsd..29). After that I chmod 600 file on the
> host, but it is still cat'able, even after move it under another name
> (i.e. no inode change?). Only after copying it to another name, client
> is unable to read it.
This one is a Linux NFS client bug. It caches the file, but forgets to
check permission on it. I will look into it if I find time.
--
H.J. Lu (hjl@gnu.org)
-
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/