| Message ID | 20180518165537.32002-4-jkbs@redhat.com |
|---|---|
| State | RFC |
| Headers | show |
| Series | ovn: Observe and check for effects of incremental processing | expand |
On Fri, May 18, 2018 at 9:55 AM, Jakub Sitnicki <jkbs@redhat.com> wrote: > > With incremental processing packet-ins handling in the ovn-controller > should not trigger logical flow processing. One such case of packet-in > handling is ovn-controller generating ICMP/TCP reset packets due to use > of ACL reject action. Ensure that logical flow processing is not > happening by watching coverage counters. > > Signed-off-by: Jakub Sitnicki <jkbs@redhat.com> > --- > tests/ovn.at | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/tests/ovn.at b/tests/ovn.at > index 4a5316510..18d2ab692 100644 > --- a/tests/ovn.at > +++ b/tests/ovn.at > @@ -9556,6 +9556,10 @@ ovn-nbctl --log acl-add sw0 from-lport 1000 "inport == \"sw0-p11\"" reject > ovn-nbctl --log acl-add sw0 from-lport 1000 "inport == \"sw0-p21\"" reject > OVS_WAIT_UNTIL([test 3 = $(ovn-sbctl lflow-list | grep 'icmp4 {' | wc -l)]) > > +for i in 1 2 3; do > + count_before="$(as hv$i ovs-appctl -t ovn-controller coverage/read-count lflow_run) $count_before" > +done > + > test_ip_packet 11 1 000000000011 000000000021 $(ip_to_hex 192 168 1 11) $(ip_to_hex 192 168 1 21) 0000 7d8d fcfe > test_ip_packet 21 2 000000000021 000000000011 $(ip_to_hex 192 168 1 21) $(ip_to_hex 192 168 1 11) 0000 7d8d fcfe > test_ip_packet 31 3 000000000031 000000000012 $(ip_to_hex 192 168 1 31) $(ip_to_hex 192 168 1 12) 0000 7d82 fcfe > @@ -9568,8 +9572,12 @@ test_tcp_syn_packet 31 3 000000000031 000000000012 $(ip_to_hex 192 168 1 31) $(i > > for i in 1 2 3; do > OVN_CHECK_PACKETS([hv$i/vif${i}1-tx.pcap], [vif${i}1.expected]) > + count_after="$(as hv$i ovs-appctl -t ovn-controller coverage/read-count lflow_run) $count_after" > done > > +# Check that hitting reject rule doesn't trigger lflow processing > +AT_CHECK([test X"$count_before" = X"$count_after"], [0]) > + > OVN_CLEANUP([hv1], [hv2], [hv3]) > AT_CLEANUP > > -- > 2.14.3 > Acked-by: Han Zhou <hzhou8@ebay.com>
diff --git a/tests/ovn.at b/tests/ovn.at index 4a5316510..18d2ab692 100644 --- a/tests/ovn.at +++ b/tests/ovn.at @@ -9556,6 +9556,10 @@ ovn-nbctl --log acl-add sw0 from-lport 1000 "inport == \"sw0-p11\"" reject ovn-nbctl --log acl-add sw0 from-lport 1000 "inport == \"sw0-p21\"" reject OVS_WAIT_UNTIL([test 3 = $(ovn-sbctl lflow-list | grep 'icmp4 {' | wc -l)]) +for i in 1 2 3; do + count_before="$(as hv$i ovs-appctl -t ovn-controller coverage/read-count lflow_run) $count_before" +done + test_ip_packet 11 1 000000000011 000000000021 $(ip_to_hex 192 168 1 11) $(ip_to_hex 192 168 1 21) 0000 7d8d fcfe test_ip_packet 21 2 000000000021 000000000011 $(ip_to_hex 192 168 1 21) $(ip_to_hex 192 168 1 11) 0000 7d8d fcfe test_ip_packet 31 3 000000000031 000000000012 $(ip_to_hex 192 168 1 31) $(ip_to_hex 192 168 1 12) 0000 7d82 fcfe @@ -9568,8 +9572,12 @@ test_tcp_syn_packet 31 3 000000000031 000000000012 $(ip_to_hex 192 168 1 31) $(i for i in 1 2 3; do OVN_CHECK_PACKETS([hv$i/vif${i}1-tx.pcap], [vif${i}1.expected]) + count_after="$(as hv$i ovs-appctl -t ovn-controller coverage/read-count lflow_run) $count_after" done +# Check that hitting reject rule doesn't trigger lflow processing +AT_CHECK([test X"$count_before" = X"$count_after"], [0]) + OVN_CLEANUP([hv1], [hv2], [hv3]) AT_CLEANUP
With incremental processing packet-ins handling in the ovn-controller should not trigger logical flow processing. One such case of packet-in handling is ovn-controller generating ICMP/TCP reset packets due to use of ACL reject action. Ensure that logical flow processing is not happening by watching coverage counters. Signed-off-by: Jakub Sitnicki <jkbs@redhat.com> --- tests/ovn.at | 8 ++++++++ 1 file changed, 8 insertions(+)