Re: /proc filesystem allows bypassing directory permissions on Linux

[psz () maths usyd edu au]

Dear Dan,

You wrote:

> >      User1 creates file with permissions 0644
> >                      User2 opens file for read access on file descriptor 4
> >      User1 chmod's directory to 0700
> >      User1 verifies no hard links to file
>
> Here's a window, during which User2 is able to create a hardlink and that will 
> remain unnoticed by User1. There's no way to perform link check and 
> conditionally do chmod in an atomic manner.

After the directory is mode 700, User2 cannot create hardlinks. There is
no window. Checking link counts anytime after the directory has been
secured, is fine.

---

Looking at older pronouncements:

http://www.securityfocus.com/archive/1/507448
> in authentic kernels /proc/<PID>/fd/<FD> are symlinks, not anything other.

http://www.securityfocus.com/archive/1/507426
> Just looked more carefully at fs/proc/base.c. That behavior is due to
> proc_fd_info() called from proc_fd_link() obtains file->f_path, that in turn
> contains the reference to the open file dentry and hence inode. That's exactly
> why those symlinks behave as hardlinks. ...
> ... I don't think this should be fixed ...

There is a self-contradiction, and a problem which might be fixed. Only
matter for debate (by opinionated people) is whether it should be fixed.

Cheers, Paul

Paul Szabo   psz () maths usyd edu au   http://www.maths.usyd.edu.au/u/psz/
School of Mathematics and Statistics   University of Sydney    Australia