Great catch. Also, very cool with sleuthing with your malloc histogram (as
discussed off-list).
Do you think it's worth mentioning in the comment describing xlate_actions()?
Acked-by: Justin Pettit <jpet...@nicira.com>
--Justin
On Aug 19, 2013, at 10:46 PM, Ben Pfaff <b...@nicira.com> wrote:
> Every xlate_actions() needs a corresponding xlate_out_uninit(), but the
> call in handle_flow_miss() lacked one. struct xlate_out has a built-in
> 256-byte actions stub, so the bug only showed up for lots of actions.
>
> Bug #19198.
> Reported-by: Ronald Lee <ronald...@vmware.com>
> Signed-off-by: Ben Pfaff <b...@nicira.com>
> ---
> This probably needs to be ported to master, but this was reported as
> an urgent bug against branch-1.11, so I've fixed it there first.
>
> ofproto/ofproto-dpif.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
> index 0f4c42c..dc89727 100644
> --- a/ofproto/ofproto-dpif.c
> +++ b/ofproto/ofproto-dpif.c
> @@ -3898,10 +3898,12 @@ handle_flow_miss(struct flow_miss *miss, struct
> flow_miss_op *ops,
> if (miss->key_fitness == ODP_FIT_TOO_LITTLE
> || !flow_miss_should_make_facet(miss, &xout.wc)) {
> handle_flow_miss_without_facet(rule, &xout, miss, ops, n_ops);
> + xlate_out_uninit(&xout);
> return;
> }
>
> facet = facet_create(miss, rule, &xout, stats);
> + xlate_out_uninit(&xout);
> stats = NULL;
> }
> handle_flow_miss_with_facet(miss, facet, now, stats, ops, n_ops);
> --
> 1.7.10.4
>
> _______________________________________________
> dev mailing list
> dev@openvswitch.org
> http://openvswitch.org/mailman/listinfo/dev
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
