Re: [REVIEW][PATCH] vfs: Lock in place mounts from more privileged users

["Serge E. Hallyn" <serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org>]

Quoting Eric W. Biederman (ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org):
> "Serge E. Hallyn" <serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org> writes:
> 
> > Quoting Serge E. Hallyn (serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org):
> >> Quoting Eric W. Biederman (ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org):
> >> > 
> >> > Serge does this patch break lxc?  I think all should be well but I want
> >> > to make certain there is not some hidden case where this fundamentaly
> >> > breaks some functionality.
> >> 
> >> I haven't yet tried.  I'll build and test a kernel today.  I'm pretty
> >> sure all the child's mounts are done after clone, so I *think* the worst
> >> case will be that the unmounting of put_old after pivot_root() will
> >> be noisy.  Will let you know.
> >> 
> >> -serge
> >
> > Just tested it - works fine.  Warns about all of the failed umounts.
> 
> Just to confirm.  Can you do a lazy umount of put_old and get rid of
> them?

Yes, it does that and it works.

> > Acked-by: Serge Hallyn <serge.hallyn-Z7WLFzj8eWMS+FvcfC7Uqw@public.gmane.org>
> >
> > ( Mind you I'm not approving of the idea of hiding mounts as a security
> > mechanisms, but I know that neither are you :)
> 
> As a security mechanism, not really.  This is more about closing a
> theoretical hole in case someone was sloppy, and doing it before user
> namespaces are too widely deployed so we avoid massive user space
> breakage.  It let's me sleep more soundly at night if I know you can't
> more access more with user namespaces that you can without user
> namespaces.

Yup.

thanks,
-serge