[ovs-dev] [PATCH] tests/bundle: test bundle action with ports up and down

Ben Pfaff blp at ovn.org
Mon Oct 3 17:10:27 UTC 2016

On Fri, Sep 30, 2016 at 06:53:37PM -0300, Thadeu Lima de Souza Cascardo wrote:
> Also, add the keyword bundle_action to all the tests in bundle.at,
> distinguishing it from OF bundles.
> It came to my attention recently that bundle_load will load 0xFFFF in case all
> the slaves are down, as bundle_execute will return OFPP_NONE.
> As I noticed this was not explicitly tested, not even for the bundle action, I
> thought it would be nice to do it as a way of documenting this behavior.
> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo at redhat.com>

Thanks.  I applied this to master.  I folded in the following
incremental, since it seemed like the behavior of bundle_load could be
more explicitly documented.

diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
index 15dbdec..16f0f7c 100644
--- a/lib/ofp-actions.c
+++ b/lib/ofp-actions.c
@@ -1200,7 +1200,8 @@ format_OUTPUT_REG(const struct ofpact_output_reg *a, struct ds *s)
  * generally take into account things like its carrier status and the results
  * of any link monitoring protocols which happen to be running on it.  In order
  * to give controllers a place-holder value, the OFPP_NONE port is always
- * considered live.
+ * considered live, that is, NXAST_BUNDLE_LOAD stores OFPP_NONE in the output
+ * register if no slave is live.
  * Some slave selection strategies require the use of a hash function, in which
  * case the 'fields' and 'basis' parameters should be populated.  The 'fields'

More information about the dev mailing list