Even though ovn-controller tries to set port_binding->chassis to NULL
every time port_binding->virtual_parent is set to NULL for bindings of
type="virtual", there's no way to enforce that an operator doesn't
manually clear the "virtual_parent" column in the Southbound database.

In such scenario ovn-controller would crash because of trying to
dereference the NULL port_binding->virtual_parent column.

Add an extra check and release "virtual" port bindings that have
"virtual_parent" NULL.

Reported-at: https://bugzilla.redhat.com/1818844
CC: Numan Siddique <nusid...@redhat.com>
Fixes: 054f4c85c413 ("Add a new logical switch port type - 'virtual'")
Signed-off-by: Dumitru Ceara <dce...@redhat.com>
---
 controller/binding.c | 26 +++++++++++++++-----------
 tests/ovn.at         | 18 ++++++++++++++++++
 2 files changed, 33 insertions(+), 11 deletions(-)

diff --git a/controller/binding.c b/controller/binding.c
index c3376e2..5ea12a8 100644
--- a/controller/binding.c
+++ b/controller/binding.c
@@ -625,22 +625,26 @@ consider_local_virtual_port(struct ovsdb_idl_index 
*sbrec_port_binding_by_name,
                             const struct sbrec_chassis *chassis_rec,
                             const struct sbrec_port_binding *binding_rec)
 {
+    if (binding_rec->virtual_parent) {
+        const struct sbrec_port_binding *parent =
+            lport_lookup_by_name(sbrec_port_binding_by_name,
+                                 binding_rec->virtual_parent);
+        if (parent && parent->chassis == chassis_rec) {
+            return;
+        }
+    }
+
     /* pinctrl module takes care of binding the ports of type 'virtual'.
      * Release such ports if their virtual parents are no longer claimed by
      * this chassis.
      */
-    const struct sbrec_port_binding *parent =
-        lport_lookup_by_name(sbrec_port_binding_by_name,
-                             binding_rec->virtual_parent);
-    if (!parent || parent->chassis != chassis_rec) {
-        VLOG_INFO("Releasing lport %s from this chassis.",
-                  binding_rec->logical_port);
-        if (binding_rec->encap) {
-            sbrec_port_binding_set_encap(binding_rec, NULL);
-        }
-        sbrec_port_binding_set_chassis(binding_rec, NULL);
-        sbrec_port_binding_set_virtual_parent(binding_rec, NULL);
+    VLOG_INFO("Releasing lport %s from this chassis.",
+              binding_rec->logical_port);
+    if (binding_rec->encap) {
+        sbrec_port_binding_set_encap(binding_rec, NULL);
     }
+    sbrec_port_binding_set_chassis(binding_rec, NULL);
+    sbrec_port_binding_set_virtual_parent(binding_rec, NULL);
 }
 
 static void
diff --git a/tests/ovn.at b/tests/ovn.at
index a846e5c..0135838 100644
--- a/tests/ovn.at
+++ b/tests/ovn.at
@@ -15010,6 +15010,24 @@ AT_CHECK([cat lflows.txt], [0], [dnl
   table=12(lr_in_arp_resolve  ), priority=100  , match=(outport == "lr0-sw0" 
&& reg0 == 10.0.0.10), action=(eth.dst = 50:54:00:00:00:03; next;)
 ])
 
+# Forcibly clear virtual_parent. ovn-controller should release the binding
+# gracefully.
+pb_uuid=$(ovn-sbctl --bare --columns _uuid find port_binding 
logical_port=sw0-vir)
+ovn-sbctl clear port_binding $pb_uuid virtual_parent
+
+OVS_WAIT_UNTIL([test x$(ovn-sbctl --bare --columns chassis find port_binding \
+logical_port=sw0-vir) = x])
+
+# From sw0-p0 resend GARP for 10.0.0.10. hv1 should reclaim sw0-vir
+# and sw0-p1 should be its virtual_parent.
+send_garp 1 1 $eth_src $eth_dst $spa $tpa
+
+OVS_WAIT_UNTIL([test x$(ovn-sbctl --bare --columns chassis find port_binding \
+logical_port=sw0-vir) = x$hv1_ch_uuid], [0], [])
+
+AT_CHECK([test x$(ovn-sbctl --bare --columns virtual_parent find port_binding \
+logical_port=sw0-vir) = xsw0-p1])
+
 # From sw0-p3 send GARP for 10.0.0.10. hv1 should claim sw0-vir
 # and sw0-p3 should be its virtual_parent.
 eth_src=505400000005
-- 
1.8.3.1

_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to