[ovs-dev] [PATCH 2/3] ovn-controller-vtep: Call binding_cleanup() before gateway_cleanup().

[Alex Wang]

On Tue, Aug 18, 2015 at 8:10 AM, Russell Bryant <rbryant at redhat.com> wrote:

> On 08/17/2015 11:29 PM, Alex Wang wrote:
> > Otherwise, binding_cleanup() will be a no-op since all related chassis
> > entries are deleted in gateway_cleanup().
> >
> > Found by inspection.
> >
> > Signed-off-by: Alex Wang <alexw at nicira.com>
> > ---
> >  ovn/controller-vtep/ovn-controller-vtep.c |    4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/ovn/controller-vtep/ovn-controller-vtep.c
> b/ovn/controller-vtep/ovn-controller-vtep.c
> > index 9a3f03b..ff25319 100644
> > --- a/ovn/controller-vtep/ovn-controller-vtep.c
> > +++ b/ovn/controller-vtep/ovn-controller-vtep.c
> > @@ -124,8 +124,8 @@ main(int argc, char *argv[])
> >
> >          /* Run all of the cleanup functions, even if one of them
> returns false.
> >           * We're done if all of them return true. */
> > -        done = gateway_cleanup(&ctx);
> > -        done = binding_cleanup(&ctx) && done;
> > +        done = binding_cleanup(&ctx);
> > +        done = gateway_cleanup(&ctx) && done;
> >          if (done) {
> >              poll_immediate_wake();
> >          }
> >
>
> This makes clenaup happen in the opposite order of the run() functions,
> which makes sense, dependency wise.  lgtm.
>
>

Yeah, I noticed this is also what we do in ovn-controller.c,

Pushed first two patches to master,

Thanks,
Alex Wang,


> Acked-by: Russell Bryant <rbryant at redhat.com>
>
> --
> Russell Bryant
>