Comment 76 for bug 769927

(In reply to comment #23)
> Here's what I don't quite get...
>
> I've bisected the problem down to the patch that switched NFS over to use
> d_automount. That doesn't really affect unmounting afaict and should really
> only be affecting things when the automount is triggered.
>
> So, it seems likely that that change is causing some sort of difference in the
> mounted filesystem. The question is what? I'll be happy to collect info from
> kernels with and without that patch to try and discern where the difference is
> -- any thoughts about what I should be collecting info on?

We should have also checked /proc/mounts in each case.