It's not easy to predict how OVS datapath flows exactly look so it's
better to avoid matching on exact ct_label masks.
E.g., on some runs of the "ECMP symmetric reply" test the datapath
flows are installed as:
ct_state(-new+est+rpl+trk).*ct_label(0x.*00000401020400000000/0xffffffffffffffff00000000)
On other runs the flows are installed as:
ct_state(-new+est+rpl+trk).*ct_label(0x.*00000401020400000000/0xffffffffffffffff00000001)
Both versions are equivalent but the test only accepts the first one.
Fixes: 4fdca656857d ("Add ECMP symmetric replies.")
Signed-off-by: Dumitru Ceara <[email protected]>
---
tests/system-ovn.at | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/system-ovn.at b/tests/system-ovn.at
index 40ba6e4..81f9332 100644
--- a/tests/system-ovn.at
+++ b/tests/system-ovn.at
@@ -5235,10 +5235,10 @@
icmp,orig=(src=172.16.0.1,dst=10.0.0.2,id=<cleared>,type=8,code=0),reply=(src=10
# Ensure datapaths show conntrack states as expected
# Like with conntrack entries, we shouldn't try to predict
# port binding tunnel keys. So omit them from expected labels.
-AT_CHECK([ovs-appctl dpctl/dump-flows | grep
'ct_state(+new-est-rpl+trk).*ct(.*label=0x.*00000401020400000000/0xffffffffffffffff00000000)'
-c], [0], [dnl
+AT_CHECK([ovs-appctl dpctl/dump-flows | grep
'ct_state(+new-est-rpl+trk).*ct(.*label=0x.*00000401020400000000/.*)' -c], [0],
[dnl
1
])
-AT_CHECK([ovs-appctl dpctl/dump-flows | grep
'ct_state(-new+est+rpl+trk).*ct_label(0x.*00000401020400000000/0xffffffffffffffff00000000)'
-c], [0], [dnl
+AT_CHECK([ovs-appctl dpctl/dump-flows | grep
'ct_state(-new+est+rpl+trk).*ct_label(0x.*00000401020400000000/.*)' -c], [0],
[dnl
1
])
--
1.8.3.1
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
