Re: Proposal: restrict link(2)

Dan Merillat (Dan@merillat.org)
Fri, 13 Dec 1996 15:23:05 -0500 (EST)


On Fri, 13 Dec 1996, James L. McGill wrote:

> Date: Fri, 13 Dec 1996 13:43:06 -0600 (CST)
> From: "James L. McGill" <fishbowl@fotd.netcomi.com>
> To: Ion Badulescu <ionut@moisil.wal.rhno.columbia.edu>
> Cc: Linux kernel mailing list <linux-kernel@vger.rutgers.edu>
> Subject: Re: Proposal: restrict link(2)
>
>
>
>
> On Fri, 13 Dec 1996, Ion Badulescu wrote:
>
> > No, you can't change the data blocks. But you _will_ change the inode by
> > incrementing the link count.
> >
> > I'm not saying that changing the link count is wrong by itself, as you can
> > also change the atime by simply reading the file..
>
> Actually, that is what saves this nonissue from the "file lifetime" argument.
>
> The owner and/or root can check the number of links and know that this has been
> done, and can find where the file is linked, and possibly by whom.

Congratulations, you just volunteered to be the guy who gets to inspect
EVERY filesystem of EVERY host to find those. You just _LOVE_ double
checking every hardlinked file to make sure it should be, right? I have
what, only about 50k files for you if you are interested.

Besides, by the time you go to look for it the damage may have been done.
Lots of bad things are done with hard links and symlinks. I'm even considering
adding the "follow symlink IFF owner matches" patch for /tmp

--Dan