[RFC] Transactional CGroup task attachment

Paul Menage menage at google.com
Fri Jul 11 19:09:42 PDT 2008


On Fri, Jul 11, 2008 at 7:03 PM, Matt Helsley <matthltc at us.ibm.com> wrote:
>
>        True, though I don't think it's as simple as just enabling lockdep. My
> understanding is you won't be able to determine if locks could ever be
> taken out of order unless all of the cgroup systems are enabled and they
> are all in the same cgroup hierarchy.

I wonder how hard it would be to extend lockdep to give a "signature"
of locking operations from point A to point B? Then you could just
compare a new subsystem with all existing subsystems without actually
running with it.

Paul


More information about the Containers mailing list