legacy netvirt: forcibly disable port security for network port

Security group applies to only VM port. Not to internal used port.
router interface port, dhc port and floating ip port.
which can be identified by its device owner.
"network:{routerxxx, dhcp, floatingip}"
It can be realized by forcibly disabling port security on those port.

With port security extension enabled, neutron disabled port security.
The extension is, however, optional, so ODL shouldn't depend on it.

Partial-bug: #7128
Change-Id: I3295ebc0e77d387cdfb80be318ef048ba9763994
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit 5da83fc96656d5be54b3451a2f221a2e780e52d4)

Checkstyle for dhcpservice-impl

Change-Id: Ifb7faabc533ce5f397e9b5b84f040c3f684785d4
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 7168 - MAC Learning from ARP to be allowed on Ext-Interfaces

An earlier fix to make sure that Neutron configured IPs are not
learned here https://git.opendaylight.org/gerrit/#/c/48054/, has
broken the ability to learn MAC Addresses on external network
via external interfaces.

With this fix, we open the gate to allow MAC learning on all
packets coming from external interface.  This is not the best
possible fix for the problem, but this tentatively allows
the openstack use-case of external network access to continue to
be functional.

Closes-Bug: 7168
Change-Id: Ief6f8bfe26111e07e280a9245a44032a8f32f09d
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Switch the NeutronFloatingIP listener to DTCL

This patch also adds a unit test to verify that the listener
registration works, and migrates to the standard Blueprint pattern of
two-phase listener creation and registration.

Change-Id: I57928b4258f5ef783dde93879136dbd9e148e820
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Checkstyle for dhcpservice-api

Change-Id: I250b7c5c890e83cb1dfe392097e35b2b8d80dfec
Signed-off-by: Sam Hague <shague@redhat.com>

Fix version warning

Change-Id: I278cdb6dd6d0fb50db54e368f1c80aa2e7fc2b42
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 6089:Add support for All ICMP code and type in SG using learn

Change-Id: Iba3bab5d04976dfbf22ef5221a33d6d0cd4602cb
Signed-off-by: hari.pr <hari.pr@hcl.com>

Add aggregator pom for commons

Change-Id: Ib999b7034a95be8a89d72d94dbf2fcd8e2df3404
Signed-off-by: Sam Hague <shague@redhat.com>

Make aggregator poms consistent

Change-Id: I24616e5c23b4d7e99697787dd2b07e6540feab6f
Signed-off-by: Sam Hague <shague@redhat.com>

Implement InterVpnLink update operation

 + An update consists in removing the InterVpnLink and create it
   again with the new data.

 + Its important to ensure that there is no trace of the old
   InterVpnLink before persisting the new one

Change-Id: I025d454e76ba8063e3bf2e98e8eb8e1c8ccb662a
Signed-off-by: Marcos López Samaniego <marcos.lopez.samaniego@ericsson.com>

Bug 7162 - legacy netvirt: null pointer exception

provider network type is optional so that it can be null. which causes
null pointer exception.
By default, legacy netvirt interprets network type as flat type.

> java.lang.NullPointerException
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.handleInterfaceDelete(SouthboundHandler.java:126)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processPortDelete(SouthboundHandler.java:190)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processPortEvent(SouthboundHandler.java:332)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processEvent(SouthboundHandler.java:274)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.dispatchEvent(EventDispatcherImpl.java:96)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.lambda$start$0(EventDispatcherImpl.java:58)[337:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_111]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_111]
>         at java.lang.Thread.run(Thread.java:745)[:1.8.0_111]

Change-Id: I9c70552ba9faa7db86ceae31996db0ff287e8c78
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit 8edd900883fcbe4e4be1a7c408268caf3b634f66)

Bug 7105: Fix learned matches for all TCP/UDP SG

Fixes learned matches for security rules allowing all
TCP/UDP to properly match the connection.
Refactor common learn implementation logic into common functions.

Change-Id: Ib94cb56cf8c026b85059f336ae042d0f1593e484
Signed-off-by: Alon Kochba <alonko@hpe.com>

Bug 7075: AlivenessMonitor skip non-neutron ports

Remove interface utilities from ElanUtils and use Genius.
Skip processing of non-neutron ports in aliveness monitor,
current code just assumes these are neutron ports and hits
an exception.

Depends-On: https://git.opendaylight.org/gerrit/#/c/48157

Change-Id: I1515b40876ad1859f7b9ae4eac384efe169a3eab
Signed-off-by: Alon Kochba <alonko@hpe.com>

BUG-7093

Do not add hw_addr configuration for integration
bridge creation when the bridge already exists on
the OVS node as this will cause the dpid on the OVS
node to change.

Change-Id: I11273e0e262c1135a351ea33e3e030116fb71cb5
Signed-off-by: Bertrand <bertrand.low@hcl.com>

Bug 7127 - legacy netvirt: null pointer exception SecurityServicesImpl

With openstack CI, the following null pointer exception occurs
getDhcpServerPort() can return null.

> 2016-11-08 09:19:17,013 | ERROR | ntDispatcherImpl | EventDispatcher | 336 - org.opendaylight.netvirt.openstack.net-virt - 1.4.0.SNAPSHOT | Exception in dispatching event NorthboundEvent [handler=NEUTRON_PORT, action=UPDATE, port=NeutronPort [portUUID=7b5ef356-7384-466b-9fd0-36366bdfd9eb, networkUUID=f1c4b809-62e3-45bc-8eec-df5ef15cb017, name=, adminStateUp=true, status=null, macAddress=fa:16:3e:be:eb:b4, fixedIPs=[Neutron_IPs{ipAddress='10.1.0.21', subnetUUID='e4b5081c-b081-47c4-af47-f5b6bb28a2de'}], deviceID=613810e5-cb2b-4c19-b7e1-f8ead6cd303b, deviceOwner=compute:None, tenantID=7cdf7d5664f143eab4ea7f3a127ff704, securityGroups=[], bindinghostID=ubuntu-xenial-osic-cloud1-s3500-5245939, bindingvnicType=normal, bindingvnicType=normal, portSecurityEnabled=false], subnet=null, router=null, routerInterface=null, floatingIP=null, network=null, loadBalancer=null, loadBalancerPool=null, loadBalancerPoolMember=null, neutronsecurityRule=null]
> java.lang.NullPointerException
>         at org.opendaylight.netvirt.openstack.netvirt.impl.SecurityServicesImpl.syncFixedSecurityGroup(SecurityServicesImpl.java:462)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.NeutronL3Adapter.processPortSecurityEnableUpdated(NeutronL3Adapter.java:936)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.NeutronL3Adapter.handleNeutronPortEvent(NeutronL3Adapter.java:440)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.PortHandler.doNeutronPortUpdated(PortHandler.java:126)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.PortHandler.processEvent(PortHandler.java:208)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.dispatchEvent(EventDispatcherImpl.java:96)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.lambda$start$0(EventDispatcherImpl.java:58)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_91]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_91]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_91]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_91]
>         at java.lang.Thread.run(Thread.java:745)[:1.8.0_91]

Change-Id: I22a50533f4e7270badefb3614147a5656e93b352
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit 8ff29e702bdc59d65c640886e2fdc7b0e3b7b7bd)

Support for IPv6 East-West Routing

This patch supports IPv6 east-west communication using OpenFlow rules.

Use-cases verified:
-------------------
1. IPv6 East-West Routing (for dual-stack and IPv6 only network)
   on an all-in-one setup as-well-as multi-node setup.
2. Above use-case in transparent mode and Stateful SG mode.

Sample setup:
-------------
1. Create a Neutron router with two networks (say network1 and network2).
2. Create an IPv6 subnet (or dual-stack) in each network.
3. Associate all the subnets to the router.
4. Spawn VMs on network1 and network2.
5. Verify that ping6 from VM1 (on network1) to VM2 (on network2) is
   successful.
6. ODL virtual router should be able to support routing between the two
   networks using OpenFlow rules.

Change-Id: I7dff4dcccf9e746656ca04cb09f99b31f585fcd8
Signed-off-by: Sridhar Gaddam <sgaddam@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 7126 - legacy netvirt: null pointer exception NeutronSubnetInterface.fromMd

With openstack CI, legacy netvirt spits the following null pointer
exception.

> 2016-11-08 09:15:31,792 | ERROR | ntDispatcherImpl | TenantNetworkManagerImpl         | 336 - org.opendaylight.netvirt.openstack.net-virt - 1.4.0.SNAPSHOT | getDHCPServerPort:getDHCPServerPort failed due to
> java.lang.NullPointerException
>         at org.opendaylight.netvirt.openstack.netvirt.translator.crud.impl.NeutronSubnetInterface.fromMd(NeutronSubnetInterface.java:178)
>         at org.opendaylight.netvirt.openstack.netvirt.translator.crud.impl.NeutronSubnetInterface.getSubnet(NeutronSubnetInterface.java:84)
>         at org.opendaylight.netvirt.openstack.netvirt.impl.SecurityServicesImpl.getDhcpServerPort(SecurityServicesImpl.java:164)
>         at org.opendaylight.netvirt.openstack.netvirt.providers.openflow13.OF13Provider.programLocalSecurityGroupRules(OF13Provider.java:1056)
>         at org.opendaylight.netvirt.openstack.netvirt.providers.openflow13.OF13Provider.programLocalRules(OF13Provider.java:814)
>         at org.opendaylight.netvirt.openstack.netvirt.providers.openflow13.OF13Provider.handleInterfaceUpdate(OF13Provider.java:1194)
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.handleInterfaceUpdate(SouthboundHandler.java:101)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processPortUpdate(SouthboundHandler.java:343)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processPortEvent(SouthboundHandler.java:329)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processEvent(SouthboundHandler.java:274)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.dispatchEvent(EventDispatcherImpl.java:96)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.lambda$start$0(EventDispatcherImpl.java:58)[336:org.opendaylight.netvirt.openstack.net-virt:1.4.0.SNAPSHOT]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_91]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_91]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_91]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_91]
>         at java.lang.Thread.run(Thread.java:745)[:1.8.0_91]

Change-Id: Id02d8e4f76c2394b6e765e7dfb92e66693051769
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit 97990d76bb26431039295f3d7d9c1b31512ee959)

BUG-7016 flows fix the bug that flows is not corrrect after the
reconnection of ovs

Fixes bug 7016
    1. do the all neutron port add when reconnect to ovs
    2. when dpip changed, change the action to add

Change-Id: I8a16c40078019096ef25b62253c47fc5bc562482
Signed-off-by: wangqianyu <wang.qianyu@zte.com.cn>

Bug 7157: Modify inter-VPN link model to enable route leaking

Change-Id: I82ee50acb9c1d9b6b6a8a548d2b484d32fe9d787
Signed-off-by: Marcos López Samaniego <marcos.lopez.samaniego@ericsson.com>

Bug 7125 - legacy: ConcurrentModificationException

With openstack devstack, ConcurrentModificationException occurs as
follows
> java.util.ConcurrentModificationException
>         at java.util.HashMap$HashIterator.nextNode(HashMap.java:1437)[:1.8.0_111]
>         at java.util.HashMap$KeyIterator.next(HashMap.java:1461)[:1.8.0_111]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.NeutronL3Adapter.updateNetworkInCleanupCache(NeutronL3Adapter.java:1625)[289:org.opendaylight.netvirt.openstack.net-vir
> t:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.NeutronL3Adapter.handleNeutronNetworkEvent(NeutronL3Adapter.java:786)[289:org.opendaylight.netvirt.openstack.net-virt:1
> .3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.NetworkHandler.doNeutronNetworkUpdated(NetworkHandler.java:100)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSH
> OT]
>         at org.opendaylight.netvirt.openstack.netvirt.NetworkHandler.processEvent(NetworkHandler.java:175)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.dispatchEvent(EventDispatcherImpl.java:96)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SN
> APSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.lambda$start$0(EventDispatcherImpl.java:58)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.S
> NAPSHOT]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_111]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_111]
>         at java.lang.Thread.run(Thread.java:745)[:1.8.0_111]

Change-Id: Icdeb4f578ef18fa34663640d78f8d0585805ae70
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit a7c57f4641ddc505262a18e9fd780c529529f5bb)

Bug 7147: VpnInterfaces not removed when InterVpnLink is removed

 + Current code is trying to remove them from CONFIGURATION, but it
   should be OPERATIONAL

Change-Id: Icaa91b5060340c4816358314b469b4a4789a5edd
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>

Bug 7124 - legacy netvirt: null pointer exception in SouthboundHandler

This patch avoids null pointer exception as follows.
With openstack tempest CI, the following exception occurs.

> java.lang.NullPointerException
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandlercd.processPortDelete(SouthboundHandler.java:184)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processPortEvent(SouthboundHandler.java:332)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.SouthboundHandler.processEvent(SouthboundHandler.java:274)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.dispatchEvent(EventDispatcherImpl.java:96)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at org.opendaylight.netvirt.openstack.netvirt.impl.EventDispatcherImpl.lambda$start$0(EventDispatcherImpl.java:58)[289:org.opendaylight.netvirt.openstack.net-virt:1.3.1.SNAPSHOT]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_111]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_111]
>         at java.lang.Thread.run(Thread.java:745)[:1.8.0_111]

Change-Id: I8549fb799ac0ae797f4b84dc8d910ef02b65a0ad
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit 2763273197ccda40b8e30bdc63ab212ce4f62903)

Bug 6786: L3VPN is not honoring VTEP add or delete in operational cloud

As part of making L3VPN honor VTEP add-delete operation , this is a first
set of change where VpnInterface can be configured and VRF entry created
even if VTEP IP on a DPN is not available.

Change-Id: If07f297be3894fe1cd5de5b614565055fa16f09e
Signed-off-by: HANAMANTAGOUD V Kandagal <hanamantagoud.v.kandagal@ericsson.com>
Signed-off-by: ehvkand <hanamantagoud.v.kandagal@ericsson.com>

Bug 6822: IVpnLink Static routes not removed on cascade

When an inter-VPN link is removed, we must ensure that there is no
NeutronRouter with an static route pointing to one of the
inter-VPN link's endpoints.

Change-Id: Ib061599fd8e1fd0a56a128d1ae1863d7c47d9170
Signed-off-by: Marcos López Samaniego <marcos.lopez.samaniego@ericsson.com>

Bug 6853 : Directly removing floatingip doesn't rmv mac entries from T19

NAT model change to accomodate and pass the mac address, uuid and subnet
id of the floating ip port for creating table 19 entries, obtained via
the created neutron port upon a floating ip create.

Subnet ID is also passed along to be used by onAddFloatingIP, thus
eliminating a neutron port lookup for the same.

These changes take care of the population, depopulation of newly added /
tweaked models and consumption of the same in NATService.
Also includes some refactoring and cleaning up.

Change-Id: I0bd8fd3937c91e885e77aeaafe4522ce4751b0fa
Signed-off-by: Abhinav Gupta <abhinav.gupta@ericsson.com>

Bug 6777 - FIB entries for RNH routed to VxLAN tunnel for flat/VLAN provider networks

Change-Id: Id0265159d3e4a9e92fc2e42f6129d7c1ff819369
Signed-off-by: Tali <tali.ben-meir@hpe.com>

BUG 7116: Change in TunnelInterfaceStateListener.java

Review [1] introduced some yang changes in genius. This review includes
relevent changes in netvirt.

[1] https://git.opendaylight.org/gerrit/#/c/46649/

Change-Id: Ide1f633bb8852562d297bf92f73b93b1faadb664
Signed-off-by: Suraj Ranjan <suraj.ranjan@ericsson.com>

Bug 6934: VpnPseudoPort flows not moved to a new DPN (II)

 + This change completes change [1]

 + When karaf is started, Cache wasn't being fed with existing data
   in MD-SAL

 [1] https://git.opendaylight.org/gerrit/#/c/46779/

Change-Id: I93d03094181093bcaa1608b3c95d6dbcfca55741
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>

Bug 6904 : ELANInstance read causing NPE on cluster reboot while sending
notifs.

ELANInstance read was causing NPE on cluster reboot while sending notifs,
leading to breakage in subnetRoute feature.

This commit depends on [1] and addresses the following:

1. On Cluster Reboot, the DataStore events are getting triggered out of
order and hence some data are not available when processing subnet and
ports addition. This fix addresses fetching elanTag from elanDatastore.
If the elan datastore is not written with the elanTag, then the elanTag
is tried to be retrieved from the IdManager.

2. Added logs indicating creation/deletion of ELAN instance.

3. Improved try-catch error handling 4. Proceeding with DjC transaction
for other operations in case this fails.

[1] https://git.opendaylight.org/gerrit/#/c/47131/

Change-Id: Icb09f4a6081b18a01ed6dc4590e695cf04fb9eb5
Signed-off-by: Abhinav Gupta <abhinav.gupta@ericsson.com>

Fixed UT failure in AclServiceTest

Change-Id: I9f4baba6218cdf9f49a791c20a51b49c904cfca3
Signed-off-by: Somashekar Byrappa <somashekar.b@altencalsoftlabs.com>

Bug 7020: Deletion issue when VM has multiple SGs with same rules

Issue: When VM has multiple SGs with same rules, single flow is created
on switch for the common rules. When one of the SG is dissociated from
the VM, that particular flow is deleted from the switch even though
there is another SG with same rules.

Fix:
- For all overlapping rules, individual flows are created on the switch
  with different priorities.
- Per SG, unique flow priority is created using IdManager. Pool range is
    1000-60000.
- Flow priorities are cached to avoid multiple lookups.
- Note, CT_STATE_NEW_PRIORITY_DROP is now set to value 50

Fix provided for stateful mode only.

Change-Id: I1d3c69883aeadb786536c3e7bf76534c1f137336
Signed-off-by: Somashekar Byrappa <somashekar.b@altencalsoftlabs.com>
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 7106: Handles static routes at InterVpnLink creation

 + When an InterVpnLink is created it is possible that there are
   some static routes pointing to any of the InterVpnLink endpoints.
   This could happen if, for example, an InterVpnLink is removed and
   created again after some time.

 + So, at InterVpnLink creation, it'll check all Neutron routers to see
   if any of them has an static route whose nexthop is one of the
   InterVpnLink endpoints. If so, the static route must be 'installed'
   in the DPN where the corresponding Endpoint has been instantiated.

 + There is a bit of refactoring. Code that was in VpnRpcServiceImpl has
   been extracted to InterVpnLinkUtils so that it can be reused.

 + InterVpnLinkCache is now created at InterVpnLinkListener creation

Change-Id: I6d53ca282f41d64c2eaa5fb1d96eec20c902ffb5
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 7086: ELAN broadcast group for VLAN network

The remote BC group for ELAN of type flat/VLAN contains buckets to
access tunnel EP but not for the provider port

Change-Id: I2d6861e86a9362b0f9ebd82f886eb05db3738da3
Signed-off-by: Olga Schukin <olga.schukin@hpe.com>

Bug 7107: RouteOrigin updated to support Local routes

 + Usage:

    BGP  - Indicates Routes Imported via BGP into ODL

    CONNECTED – Every subnet (for example a VPN) has a route directly connected
            for that subnet. Same description than Cisco (all /24 for us)

    STATIC -  The static routes introduced by the OpenStack command “extra routes”

    SELFIMPORTED - Indicates this Route is replayed route on a new BGP VPN,
            which is importing RT from another BGPVPN which originally owns
            this route

    LOCAL - https://supportforums.cisco.com/document/63281/local-host-routes-routing-table
            (basically /32 host routes).

    INTERVPNLINK – Indicates Routes replayed by InterVPNLink due to
            Route Leaking. Inter-VPN link is configurable for leaking BGP,
            connected and/or static routes. LOCAL routes are not leaked,
            this is why this change is introduced.

 + RouteOrigin is intended to be in line with:

   http://study-ccna.com/connected-static-dynamic-routes/ and
   https://supportforums.cisco.com/document/63281/local-host-routes-routing-table

Change-Id: I7afdfd7d2db8d36c62ad8ee1279daa33d04f6932
Signed-off-by: Marcos López Samaniego <marcos.lopez.samaniego@ericsson.com>

Bug 7101: Restrict ARP to only learn non-neutron IPs

Current ARP Handling code was trying to form
learned vpnportiptoport entries for configured
neutron ports incorrectly.  This is observed to
be happening primarily during the swap window from
internal to external VPN and vice-versa.

The fix attempts to make sure that the VPNInterface
we are servicing ARP for is indeed the same as configured by the
tenant.

Change-Id: I17bb53a2997e77a9ba7ce13284525ac403fd2bac
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Bug 7120 : NAT Support For GRE TEP add/del is missing

Support has been added in the NAT for the GRE TEP changes. Below are the
enhancements :

1) When GRE TEP is deleted :

a) SNAT : If the corresponding DPN was the NAPT switch, then a new NAPT
switch is
re-elected among the other DPNs and rest of the non-NAPT switches point to
the new NAPT switch. The routes to the External Fixed IPs (SNAT IPs) are
updated with the new NAPT switch address.

b) DNAT : If the corressponding DPN was hosting VMs with the Floating IPs,
then
the routes to those FIPs are withdrawn from the BGP and removed from the
controller FIB.

2) When the GRE TEP is added :

a) SNAT : The correponding DPN will be delegated as the non NAPT switch
and made to point to the exsiting NAPT switch. Routes are not altered.

b) DNAT : If the corresponding DPN was hosting VMs with the Floating IPs,
then routes to those FIPs are advertised to the BGP and installed in the
controller FIB.

Change-Id: I0cb968b7beb9ec294cc4bafb6539e2e6cdd1c541
Signed-off-by: Satish Dutt <satishd@altencalsoftlabs.com>

Fixes Bug 7076 SSH between vm in different network on same compute is
blocked even with an allow rule.

the conntrack +trk rule is removed in egress pipeline. Now the packets
will be always send to netfilter  for checking the status of the
connection in a zone.

the ct_state check is removed from table 40 and 251, and packet will be
send to netfilter without checking the ct_state.

fixed the UT test cases, no new UT added as this fix is not for a
particular scenario and exsiting UT will cover the change.

Change the priority on +trk+inv as the invalid packet will be always
dropped, so it is not required to check all the +trk+new rules.

Change-Id: I61f347ef537cba4d144ca2e9f8677434ca200cbc
Signed-off-by: Aswin Suryanarayanan <asuryana@redhat.com>

Fix one of the Netvirt IT test cases failure.

   - Add logs to debug the rest of the test cases failures.

Change-Id: I0fca6790b70f921d6f9f552835d809d34f3f63b4
Signed-off-by: vinoth <m.vinoth@hcl.com>

Listens to Network-topology nodes instead of inventory ones

 + Listening from opendaylight-inventory nodes changes fires too many
   updates that are not needed for the purpose of
   InterVpnLinkNodeListener.

 + So this listener will be now notified for changes in 'flow' nodes under
   network-topology.

Change-Id: I97e4e977678e9dd3bc7c86617a2838c2d4caa100
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>

Bug 7034 : Replace all write_actions by apply_actions.

write_actions is used to aggregate action (in actions set) to be apllied once there is no more goto table instruction.
In our case there is no need of this, only complicates the debugability and causing bugs.

Related review:
https://git.opendaylight.org/gerrit/#/c/46538/

Change-Id: I3324e25ed30b876d2ee0fea30c2c53b24cedf528
Signed-off-by: Slava Radune <slava.radune@hpe.com>

small fixes related to Bug 7031

- write an empty nexthop ip address
- check if the primary adjacency before its using
- increase the flow priority

Change-Id: Ifb6646553ef60a6de91346a021d8b2ec0e9d75b0
Signed-off-by: Olga Schukin <olga.schukin@hpe.com>

Bug 7091 : When Primary NAPT switch goes down, NAPT switch re-election is
not happening.

The root cause of this issue was the change in the representation of the
ietf-interfaces:interfaces-state model.

Since NAT models depends upon the ietf-interfaces:interfaces-state model,
NAT HA functionality of re-electing the NAPT switch never got triggered
and the functionality got broken. As per the new representation of the
'ietf-interfaces:interfaces-state' model, 'oper-status' attribute value is
'unknown' if the corressponding DPN goes down.

update() of the NatInterfaceStateChangeListner is implemented to handle this change for
clearing or populating the 'dpn-vpninterfaces-list' in the NAT model
'odl-l3vpn:neutron-router-dpns' used by the NAT HA code.

Change-Id: I533c9c1c5d32ab5967cd21e7b0f203978b074703
Signed-off-by: Satish Dutt <satishd@altencalsoftlabs.com>

Bug 7074
Fix wrong validation check in dynamic tunnel creation logic

Change-Id: I176e489b1aa4b54a905c0c9d86f58d9029dc5168
Signed-off-by: Ravit Peretz <ravit.peretz@hpe.com>

Bug 7055: Interface removals donot keep Op DS and Cfg DS consistent.

Southbound driven interface removals (if northbound
port-delete gets delayed), fail to keep the
operational and config DS in sync.

This fix makes sure the missing operTxn submission
happens and it happens before the config DS submission.

Change-Id: I3a1e404c5eee12c0d69dea4eddd7d202efb72d69
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Bug 7084: Fix for IP is  not assigned in single OpenStack node while creating more than one VM at a time.

Change-Id: I848a6a864a9898f5fdb98404970e026f90ffda36
Signed-off-by: bharathi.m <bharathi.m@hcl.com>

BUG 6940 - Avoid TZ subnet per neutron subnet

Before this patch, each neutron subnet was being mapped to a transport
zone subnet. This is not correct, as the transport zone subnets relate
to the provider subnet each TEP belongs to and not to overlay subnets.
The observed problem is that the same TEP could be configured multiple
times on different transport zone subnets. This would result on ITM
getting confused and thinking that some TEPs are not longer needed and
removing them.

Since at this point there is not particular reason to group the TEPs by
subnets, this change proposes to configure the TEPs on the 0.0.0.0/0
subnet.

Change-Id: Id869fae5417aa9bff292b9bb1e738e05c7ea9bd1
Signed-off-by: Jaime Caamaño Ruiz <jaime.caamano.ruiz@ericsson.com>

Bug 7045: ACL: Default flows are not programmed in Cluster environment

AclNodeListener (which is a DataTreeListener) had ACL service entity owner
check. Because of this check, default flows were not pushed to the switch
in clustered environment. As a fix for the issue, this check has been
removed now

Change-Id: Ia33d22dbf65db28c4963a9375a341702af2c5503
Signed-off-by: Shashidhar Raja <shashidharr@altencalsoftlabs.com>

Fix for ACL UT failure

+ In Genius, MetaDataUtil.METADATA_MASK_LPORT_TAG was updated recently.
Hence updated the same in ACL UT.

Change-Id: I059a9036481b5264e99bc00a8987e7f4cfea4e1f
Signed-off-by: Somashekar Byrappa <somashekar.b@altencalsoftlabs.com>
(cherry picked from commit 1feeb9a2604ac8d0c645358a691bd9173b93de46)

Bug 6992 - legacy: ignore IPv6 router interface

Since NeutronL3Adapter doesn't know ipv6, it should ignore
ipv6 router interface. Especially floating ip doesn't work
depending on the situation.

Neutron Router can be attached with both of IPv6 only port and IPv4 only
port. IPv6 only port should be ignored when programming for floatingip.
Otherwise networkIdToRouterMacCache can populated with mac address of
IPv6 only port, and programFlowsForNeutronRouterInterface() may program
flow with IPv4 address with mac address of IPv6 port. It should use
mac address of the IPv4 port. Otherwise floatingip is broken.

Change-Id: I87ec4e2537666f5db6826ca93be8be1793157ec7
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
(cherry picked from commit a649a85da5d16d917a6c6bf3982c3a93e22de228)

Disable SG IT test until learn test is included

Change-Id: I6a64dd4f02cfa65f2145d2c687189aa0dd34a1fe
Signed-off-by: Sam Hague <shague@redhat.com>

Do not add br-int when manually deleted

Change-Id: I87f9acad3c42d0fce81c411917b331052a99a913
Signed-off-by: Sam Hague <shague@redhat.com>
(cherry picked from commit 193ddaf380fe58c92816f2047ef86e17a94c645f)

IT - L3 tests

Added new tests to validate L3 connectivity:
- create 2 different networks with 2 ports.
- validate that ping between the ports is not working.
- create a router and connect the networks to it.
- validate that ping is working.

Change-Id: Ie9f5fa9b4cb10447fd15b313cfbee47ab29e718f
Signed-off-by: Oded Shvartz <oshvartz@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>
Author: Sam Hague <shague@redhat.com>

Do not log frequest NeutronHostConfig updates

Change-Id: Ic56df1bee3d2cc2c2dbb175ff974a7f4c99a7846
Signed-off-by: Sam Hague <shague@redhat.com>
(cherry picked from commit debfd73ebe081d605fda943f8b678008986110d4)

Bumping versions by 0.0.1 for next dev cycle

Change-Id: Ib7e99051b18f96f7151d43427277824329a52eef
Signed-off-by: Anil Belur <abelur@linuxfoundation.org>

Fix for Bug 7059

Included Offset in the table numbers for resubmit and the learn rules

Change-Id: Ibaed3da431de62d1583d9463cd5cdbc06c9d0720
Signed-off-by: gvrangan <venkatrangang@hcl.com>

Bug 7024: When router is associated to L3VPN , VRF entry creations takes
long time

When router is associated to L3VPN and dis-associated it takes long time
for VRF entries to get created and also for table=21 entries to appear on
DPN.

When swap from internal-to-external VPN or vice-versa happens,
VpnInterfaceManager waits on notify task for 50sec eventhough FIB has
already cleaned up. This is due to notification not fired from VpnInterface
Oplistner.

Change-Id: Ic16febd1f322f5c79d17e428a8f95133e7fef227
Signed-off-by: HANAMANTAGOUD V Kandagal <hanamantagoud.v.kandagal@ericsson.com>

Bug 6089: Fix the wrong implementation for ICMPV6

  * As of now learn supports only ICMPV4 code(0) and types(8,13,15,17)
  * reverting the ICMPV6 changes

Change-Id: I621eccf7707b5a9cba835d35df3446568158efd5
Signed-off-by: hari.pr <hari.pr@hcl.com>

Bug 7031: Implement ping responder for router interfaces

Trello task:
https://trello.com/c/z19VUYoU/151-ping-responder-for-router-interfaces-ipv4

Required for several tempest tests that ping the router interface.

This change depends on the following reviews:
[1] https://git.opendaylight.org/gerrit/45192
[2] https://git.opendaylight.org/gerrit/45632
[3] https://git.opendaylight.org/gerrit/46219

Change-Id: Iac9ad2d70fecba823630281142b632757c3a48c2
Signed-off-by: Olga <olga.schukin@hpe.com>
Signed-off-by: Olga Schukin <olga.schukin@hpe.com>
Author: Sam Hague <shague@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 6476 : After configuring NAPT, table 26 and table 46 are not programmed

Issue was observed when interfaces were added to router, after the VM was
spawned, due to the timing between two threads. Addressed this issue

Change-Id: I5171bd98df28a4fe59daace7c1a2ad444e1bb1e0
Signed-off-by: Satish Dutt <satishd@altencalsoftlabs.com>
Author: Sam Hague <shague@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Changed the AsyncDataChangeListenerBase to
AsyncDataTreeChangeListenerBase in the NAT reated files

Addressed the review comments

Change-Id: I9fa06d9d230839ab8b812470469e14f247b7d8f5
Signed-off-by: Satish Dutt <satishd@altencalsoftlabs.com>
Author: Sam Hague <shague@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Fix missing init for VpnPseudoPortListener

2016-10-21 17:21:27,920 | ERROR | rint Extender: 1 | BlueprintContainerImpl           | 15 - org.apache.aries.blueprint.core - 1.6.1 | Unable to start blueprint container for bundle org.opendaylight.netvirt.cloud-servicechain-impl/0.4.0.SNAPSHOT
org.osgi.service.blueprint.container.ComponentDefinitionException: Component 'vpnPseudoPortListener' does not have init-method: init
    at org.apache.aries.blueprint.container.BeanRecipe.getInitMethod(BeanRecipe.java:645)[15:org.apache.aries.blueprint.core:1.6.1]
        at org.apache.aries.blueprint.container.BeanRecipe.internalCreate2(BeanRecipe.java:833)[15:org.apache.aries.blueprint.core:1.6.1]
            at org.apache.aries.blueprint.container.BeanRecipe.internalCreate(BeanRecipe.java:811)[15:org.apache.aries.blueprint.core:1.6.1]
                at org.apache.aries.blueprint.di.AbstractRecipe$1.call(AbstractRecipe.java:79)[15:org.apache.aries.blueprint.core:1.6.1]
                    at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_102]
                        at org.apache.aries.blueprint.di.AbstractRecipe.create(AbstractRecipe.java:88)[15:org.apache.aries.blueprint.core:1.6.1]
                            at org.apache.aries.blueprint.container.BlueprintRepository.createInstances(BlueprintRepository.java:255)[15:org.apache.aries.blueprint.core:1.6.1]
                                at org.apache.aries.blueprint.container.BlueprintRepository.createAll(BlueprintRepository.java:186)[15:org.apache.aries.blueprint.core:1.6.1]
                                    at org.apache.aries.blueprint.container.BlueprintContainerImpl.instantiateEagerComponents(BlueprintContainerImpl.java:724)[15:org.apache.aries.blueprint.core:1.6.1]
                                        at org.apache.aries.blueprint.container.BlueprintContainerImpl.doRun(BlueprintContainerImpl.java:411)[15:org.apache.aries.blueprint.core:1.6.1]
                                            at org.apache.aries.blueprint.container.BlueprintContainerImpl.run(BlueprintContainerImpl.java:276)[15:org.apache.aries.blueprint.core:1.6.1]
                                                at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_102]
                                                    at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_102]
                                                        at org.apache.aries.blueprint.container.ExecutorServiceWrapper.run(ExecutorServiceWrapper.java:106)[15:org.apache.aries.blueprint.core:1.6.1]
                                                            at org.apache.aries.blueprint.utils.threading.impl.DiscardableRunnable.run(DiscardableRunnable.java:48)[15:org.apache.aries.blueprint.core:1.6.1]
                                                                at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)[:1.8.0_102]
                                                                    at java.util.concurrent.FutureTask.run(FutureTask.java:266)[:1.8.0_102]
                                                                        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)[:1.8.0_102]
                                                                            at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)[:1.8.0_102]
                                                                                at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)[:1.8.0_102]
                                                                                    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)[:1.8.0_102]
                                                                                        at java.lang.Thread.run(Thread.java:745)[:1.8.0_102]

Change-Id: I94825cd3e8ec6a33cc17a4ac76e26663ba1afe22
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 6717 - Output to external network group entry is not installed on NAPT FIB table for new DPN

Change-Id: Ic76eaba6b6840315513830cbf846ebc0cef42a7a
Signed-off-by: Tali <tali.ben-meir@hpe.com>
Signed-off-by: Sam Hague <shague@redhat.com>
Author: Sam Hague <shague@redhat.com>

Bug 6831: support for l3 directly connected subnet
After the fix only  unique mac values will be stored in the vpn interface adjacency.
This values will be used for the group programming.
No duplicate groups will be created.

Change-Id: I73c0e726c28e258eed3f3559e114003887a391a6
Signed-off-by: Konsta Pozdeev <konsta.pozdeev@hpe.com>

Bug 6778 - VPN interface for external port is deleted when clearing router gw interface

This patch fixes several external VPN interface issues
1) VPN interface will be created for each newly created external ELAN interface
and deleted on ELAN interface deletion
2) External VPN interfaces + VPN instance will be deleted only after ext-net was deleted

Change-Id: Ie358c214afcea1f894db6c46fe9827e637f52863
Signed-off-by: Tali <tali.ben-meir@hpe.com>
Signed-off-by: Sam Hague <shague@redhat.com>
Author: Sam Hague <shague@redhat.com>
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Bug-6395: Fixed the Problems in using ODL and neutron-l3-agent in Openstack

  * Added following flow is missed out,
      cookie=0x0, duration=7451.336s, table=100, n_packets=147, n_bytes=6230, idle_age=6506,
      priority=4,in_port=5,dl_src=fa:16:3e:74:a9:2e actions=output:3

Change-Id: I240db06a703124320b5658685226d4d2c858061a
Signed-off-by: vinoth <m.vinoth@hcl.com>

Bug 6089:Fix for TCP/UDP and ICMP communication between VM's using learn Action according to SG

    *Introduced new tables(31,39) in the netvirt pipeline.
    *Table 30 default pipeline rule is modified to forward the packets to table 31.
    *Table 31 is used for double resubmit action [actions=resubmit(,39),resubmit(,40)].
    *Table 39 is used for learned flows.
    *Flows matching in Table 39 will load reg6=0x1.
    *table 40 and 90 will use reg6 value to forward the packets to next pipeline.
    *Learn rule will be created in table 40 and 90
    *once user apply ingress/egress on VM instance.

Change-Id: Idb71bbf9ae53d53db1197622eef02047f0e86e18
Signed-off-by: hari.pr <hari.pr@hcl.com>

InterVpnLink cache

 + This is only a first step, there are many other MDSAL readings
   related to InterVpnLink that will be replaced later on

 + Since InterVpnLink is rarely modified (almost never) and
   very frequently queried, it is the perfect candidate for being
   cached.

 + This cache is placed in vpnmanager-api and accessed from
   vpnmanager, neutronvpn, fibmanager and bgpmanager. However
   it is only feeded from vpnmanager.

 + There are 2 new Clustered DCN listeners in charge of feeding
   the cache, one listener for InterVpnLink and another for
   InterVpnLinkState objects.

Change-Id: I2582784650bd8fafbd311c103669928e5bd2fb97
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>

Bug 6934: VpnPseudoPort flows not moved to a new DPN

 + when a VM is moved to a new DPN, VpnPseudoPort flows that
   moves traffic from VPN to SCF (tables 20 and 17) were not being
   installed on the new DPN

Change-Id: I8fb6daee68ec6203397c4096259a531401b109df
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>

Bug 6863 - Router interfaces incorrectly include network interfaces

When a network is associated after a router is associated
to a given VPN, interfaces on that network gets queued
as though they are on the router.

This change fixes that problem by making sure that
NeutronRouterInterfacesMap is populated/deleted only
with the correct interfaces available on that router
alone regardless of other networks that may be present
on the same VPN as the router.

Change-Id: Ib436e608f7fee398f1cc51b5ae24ac4c7de3d87a
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

AclServiceTest with immutables.org "depluralize" option

Change-Id: I3bde1f5d5373d4b9fc64f5b530f7820d5780acb8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Cleanup: remove unnecessary boxing/unboxing

Change-Id: I5b5cbb5879d12f3cc285624e489cc0544af8080c
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Cleanup: use Java 8 lambdas

Where appropriate, use Java 8 lambdas instead of old-style
predicates. Simplify surrounding code where relevant.

Change-Id: Id39da6bcdcae15ff2fed8cdd2327f6e25d79fcc3
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bug 6482: ERROR Log Observations - CSIT (Boron-Legacy)

   * This exception is from CSIT logs when external interface with matching Neutronport found.
   * While users verifying CSIT logs sometimes misunderstood of this error.
   * It's better we can change this error message to warn message.

Change-Id: Ia7fe5606cd18c17c4ee2dd8a4f4feb21c41a524a
Signed-off-by: priya.ramasubbu <priya.ramasubbu@hcl.com>

Fixes Bug 6909
ACLs TCP/UDP port ranges for the case of all ports 1-65535) should not
use port masking at all

Added check to handle all port case.

Patch 2 Added UT.

Change-Id: Ifb972c338c9cd27885a58bcff9d11969bb62cd2a
Signed-off-by: Aswin Suryanarayanan <asuryana@redhat.com>

fix learn security groups

Change-Id: I9377e5cc4c2b7623baef922e3dc1bead1cf3d2f5
Signed-off-by: Shlomi <shlomi.alfasi@hpe.com>

Arp cache feature changes

Change the implementation of Arp Cache feature so that
the Arp entries learnt for the VMs are removed only when
those VMs are no longer accessible to the gateway.

1) Remove the old implementation where as soon as a VM's mac
is learnt, it is enqueued in a queue with a timer and this entry
is removed after the timer expires. After the removal from queue,
the mac and FIB entries of the VM were removed.

2) When a VM's arp is learnt, use the Aliveness Monitor module to
monitor the VM's by sending out ARP packets from gateway to the
VM periodically. The aliveness monitor module should listen to the
VpnPortIpData which would contain the ports present in the vpn.

3) Unless the monitor event "down" is received,ie,the VM is no
longer accessible from the gateway, the Mac entries for the VM are
to be retained.

4) When the monitor event "down" is received, the mac entries for
that VM and their FIB entries are to be removed.

Change-Id: I4c29e4390a1c5be0609e8ee1d682728ab0762588
Signed-off-by: Gobinath <gobinath@ericsson.com>

bug 6643 fixed broken l2gw functionality

some parts of the code were commented during blueprint migration.
fixed the same.

Change-Id: I50601948a97dc66a60c87d943458a87dab12b43d
Signed-off-by: K.V Suneelu Verma <k.v.suneelu.verma@ericsson.com>

Bug 6816: NAT breakage fix for GRE provider type

Recent changes in NAT to support transparent vpn caused issues
for existing NAT usecase which works on GRE provider type. Fix is
provided so that provider type is checked when external network is
created.

https://git.opendaylight.org/gerrit/#/c/45133/3
https://git.opendaylight.org/gerrit/#/c/43130/

Change-Id: I0be43018dd681572496ddf7c9fa29dbfc8da2dcf
Signed-off-by: Yugandhar Sarraju <yugandhar.s@altencalsoftlabs.com>

Bug 6831: Retain subnetroute with l3 directly-connected subnet

The following review :

https://git.opendaylight.org/gerrit/#/c/46292/

was reverted via:
https://git.opendaylight.org/gerrit/#/c/46575/

The revert was done in order to maintain sanity of the
subnetroute feature which enables auto-discovery of
Non-Neutron IP addresses on an enterprise network by using
the datacenter as the proxy.

The revert however resulted in use-case breakage where rewrite of
MACAddresses held behind a directly connected subnet ceased
to happen.

With this fix, we bring back the use-case of rewrite of mac-addresses
for hosts behind a directly connected l3 subnet.  In addition to that
we attempt to maintain a working subnetroute functionality.

Change-Id: I89a61396d62e38cf9756cbd0bdb2b845b5e8e44f
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Bug 6843 :  NPE in router-add leading to failure of router related cases

The commit https://git.opendaylight.org/gerrit/#/c/45511/1 introduced an
NPE due to parameter original being null

Have added required additional null checks.

Also, tweaked the snatSettingChanged() and externalFixedIpsChanged() to
truly indicate a change and not only upon a router update when external
network is present as was the earlier behaviour.

Change-Id: Ibfd25bc0e2106d450b9045c92e887464447e2d91
Signed-off-by: Abhinav Gupta <abhinav.gupta@ericsson.com>

Bug 6779 -After a Cluster Reboot, 10 VPNintfs seen

After a Cluster Reboot, only 10 VPN Interfaces are configured
out of 1500 VPN Interfaces.

Upon cluster reboot , VpnInterface add may fail due to
vpnInstance unavailability. VpnInstance creation might lag
VpnInterface add event.

So we have introduced a synchronization mechanism for
VpnInterfaceManager to synchronize with VPNInstance
creation driven by VpnManager.  The same concept applies
to the VpnSubnetRouteHandler engine.

With this fix VpnInterfaceManager/VpnSubnetRouteHandler
will wait for 90 seconds for VpnManager to provide complete
VpnInstance information before proceeding to plumb the
VpnInterfaces/SubnetRoutes that are part of that VPNInstance.

These waits will be removed in Carbon as we refactor and will
build a parent first processing logic around the
VpnInstanceListener.

Change-Id: I912fda5aef4b4c50d897cd293dee6f49c429eea6
Signed-off-by: Vivekanandan Narasimhan <n.vivekanandan@ericsson.com>

Bug 6824 - floating IP rules deleted upon unrelated neutron port delete

Change-Id: Id3f9c3df057efaddb34628c3df389485a5d99584
Signed-off-by: Abhinav Gupta <abhinav.gupta@ericsson.com>

Increase AclServiceTest coverage significantly (from 66% to 84%)

Incl. new DataBroker Tx infra utils (to be moved elsewhere later)

Incl. (first use of) http://immutables.github.io - very cool.

Change-Id: Iaf649800d3da5aa2081e129bb03ecdf604e06075
Signed-off-by: Aswin Suryanarayanan <asuryana@redhat.com>
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Bug 6923 - sfc-translation-layer : OVS data path
locator options (nsp,nsi,nshc*) are not required.

Apart form the netvirt expect data path locator name
to be vxgpe. This hardcoded dependency needs to be removed.
Following bug tracks the issue
https://bugs.opendaylight.org/show_bug.cgi?id=6924

Change-Id: Ic08cd65a547a93d1e878fcf8bce7d5f66a894b9b
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>

Bug 6922 - sfc-translation-layer : Do not explictly set RSP name

SFC translation layer currently set the RSP name to SFP name
and that results in flow installation with two different nsp
flows. Given that RSP RPC returns the RSP name, translation
layer should use that name.

Change-Id: I23a88866c1348575b9a40911dbb573e6ef0f5344
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>

Bug 6921 - SFC-Translation-Layer : Skip acl classifier
write before chain creation

Currently SFC translation layer, converts the networking-sfc
flow classifier to acl classifier and write it to data store.
The way current net-virt works, it should be written once the
chain is created. In the current translation layer implementation,
 it writes both the time and second write become no-op and
netvirt don't install the flow classifier rules.

Change-Id: I92e5065492e4c269c0b0d5f9f23afde2de516335
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>

Bug-6395: Fixed the Problems in using ODL and neutron-l3-agent in Openstack

  * Now we can use ODL for L2 and neutron-l3-agent for all L3. This can help covering more use cases with ODL integration.
  * invoke patch port creation during set-manager if ovsdb.l3.multiple.network.enabled flag enabled.
  * Create patch-ports in br-int and br-ex
  * create flow entries in br-int(table 100) to handle packets coming through patch port via external bridge.
  * create flows in external bridge to push vlan based on vlan provider network segmentation_id

Change-Id: I0edf0751a72df97f77576b757b2c2145a5406543
Signed-off-by: vinoth <m.vinoth@hcl.com>

Bug 6920 : Fix for ACL portSecurityUpdate to work with DjC + listed fixes

1. Fixes conflicting modification exceptions at interfaces/interface level
due to portVifType update and securityGroupUpdate writing at same time.

2. Includes fix for Lock API for sync in NeutronVpnUtils causing
concurrent modf exception : Bug 6902 :
https://git.opendaylight.org/gerrit/#/c/46663/

3. Synchronize on writes to router-interfaces map, not using DjC

Change-Id: I74bf6dd7bd80539e483bdf233a6d91ea39031190
Signed-off-by: Abhinav Gupta <abhinav.gupta@ericsson.com>

De-static-ify aclservice utility classes methods and fields

Don't ever use static methods and fields, but instead just use
@Singleton and proper dependency injection (DI).  Ask me if you are
unclear how to do this right.

This change isn't just for fun, but will avoid an ugly work-around we
needed to do in the new component tests, to reset a cache hidden in a
static via a (temporary, hack, yes) AclDataUtil.clearCaches(); in
AclServiceTest in https://git.opendaylight.org/gerrit/#/c/46206/

Change-Id: I1d1ba36739784a08001d66faa6f9641551a9036e
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

aclservice end-to-end test, with a bunch of cool new patterns

Depends-On: https://git.opendaylight.org/gerrit/#/c/46554/
Depends-On: https://git.opendaylight.org/gerrit/#/c/46557/

This originally increased coverage of aclservice from 53% as of de3d70d
to 65% (but things have moved a bit since this work was started). The
new approach introduced should make it easy to push it up further.

Meant also as an example for replication elsewhere; illustrates:

* manual Xtend objects for Flows
* generated Xtend objects by XtendBeanGenerator
* AssertBeans helper for readable test object comparison
* Mockito with CallsRealOrExceptionAnswer etc. (formerly "Mikito")
* Wiring DI best practice pattern, based on Guice

Change-Id: Icfb0a77ebfe807d30ea591d2997ef76718d2d0d1
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
(cherry picked from commit 50dc4ed226a4a18ec151c93a6e966a974a680bb8)

Fix & un-@Ignore AclServiceTest

It turns out that c/45261 re. "@Override init() instead @PostConstruct
start()" broke AclServiceTest (c/42109); somehow Mycila Guice
Jsr250Module interprets @PostConstruct in parent class, or protected VS
public, differently than Apache Aries Blueprint impl.

This fixes it by putting @PostConstruct back, for now.  I'll separately
investigate a cleaner fix (probably going either in Mycila Guice, Genius
or infrautils; not netvirt).

This problem was not seen pre-merge, and thus broke master, because we
are in the unhealthy habit of merging a bunch of changes without
re-re-rebasing EVERY change after EVERY merge.  While that certainly
adds some overhead for commiters and build infra, doing that would lead
to more stability in such cases, as well as avoid Guitar Hero git log
<https://twitter.com/henryhoffman/status/694184106440200192>, which in
turn would make it easier to locally try git revert and git bisect to
find out which commit cause what.

Change-Id: Ife74e3b920cca1fdf43fdf3aedc6401723050c46
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Upgrade AssertBeans version, intro. nicer syntax w.o. build()

Also incl. some general clean-up: now all code in *-impl instead of some
in *-api, as that is probably much cleaner, and simpler to follow when
reading.

The 3 classes that are in test.infra will eventually move somewhere else
(but let's anyway get this in as proposed here ASAP already).

Change-Id: Iaf136a88d1f6740a56159efc3c5349348b9516b0
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Clearer failure message in AclServiceTest if no flows are created

Change-Id: I6832dd4af90865bb0227bb097271af8b25e932b8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Signed-off-by: Sam Hague <shague@redhat.com>

Remove unneeded alivenessmonitor-xml css dependency

Change-Id: I325139a6d9075ab63439b65d18f54b8b876e32f7
Signed-off-by: Sam Hague <shague@redhat.com>

Bug-6474 : Fixed the issue when using ODL with VXLAN Gateway

    Problem: The ARP request from a VLAN host can reach the VM
    but ARP reply is dropped because it is an unknown unicast.

    Solution: Add metaflow in OVS that creates learned flow
    for convert the source MAC to Dest MAC and send the packet
    back to the tunnel where the packet was received.

    To enable this feature, please add the following line
    into the file "etc/custom.properties".

        ovsdb.remotemac.learn.enabled=yes

Change-Id: I05b3a0fa63d71ea733ad0de53fc1ad37ed20f785
Signed-off-by: Hideyuki Tai <Hideyuki.Tai@necam.com>

Add clear ping status

Change-Id: Id51ebc3d033389ff229a4ec864ac12bef7b5faa1
Signed-off-by: Sam Hague <shague@redhat.com>

Fix BROKEN aclservice listeners

This fixes it by putting @PostConstruct back.

This is the (partial) back-patch of https://git.opendaylight.org/gerrit/#/c/45966/
from master to stable/boron.  (While that change was initially thought to be
a temporary fix on master, it ultimately turned out to be the right and only
required fix; more technical background about this is meanwhile documented on
https://wiki.opendaylight.org/view/BestPractices/DI_Guidelines.)

This is only a partial selective cherry-pick, because the (un-@Ingore)
in AclServiceTest has been excluded here, because there is no AclServiceTest
working back-patch on stable/boron, yet.

see https://lists.opendaylight.org/pipermail/netvirt-dev/2016-October/001739.html

Change-Id: Ife74e3b920cca1fdf43fdf3aedc6401723050c46
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Fixes logging exceptions, plus few formatting changes

 + Logging exceptions now comply with [1]

 + Adds some 'final' qualifiers

 + Also adds some missing braces for 'if' blocks

 + Removes some unused imports

 [1] https://wiki.opendaylight.org/view/BestPractices/Logging_Best_Practices

Change-Id: Idf860e0e8b87dfa740e5d0f3d6e1dfe14cce564d
Signed-off-by: Miguel Perez <francisco.miguel.perez@ericsson.com>
(cherry picked from commit 01a2208087ea2f547c95ee79a816e7dfdfc39d09)

Drop Maven prerequisite

The Maven versions are enforced by odlparent, projects no longer need
to care about them (unless a particular project-specific plugin
requirement crops up).

Change-Id: Iffc2f96e29ce0b5e5a1b5f91ff7280895f15eb16
Signed-off-by: Stephen Kitt <skitt@redhat.com>

ipv6: Use versions from odlparent

Change-Id: If055af1a68e53edabf9a90c04aac9df90a4325ba
Signed-off-by: Sam Hague <shague@redhat.com>

Revert "Bug 6831: support for l3 directly connected subnet"

This reverts commit 53edd46eb630dcc069c8378e2800b326f04c4239.

Change-Id: I9ad7cac6311ee2515528edf34f112588c19b8d65
Signed-off-by: Sam Hague <shague@redhat.com>

Remove duplicate lockmanager bean

Change-Id: Ib31218a64410d3180b1e4d01912410c6ea738b5c
Signed-off-by: Sam Hague <shague@redhat.com>

Revert "Bug 6089: Fix for communication between VM's according to SG."

This reverts commit e130135709d36e3edb5709e4a68e774200b16af7.

Change-Id: I170e2b01fa7c22477d2627df06a08f9cc7565560
Signed-off-by: gvrangan <venkatrangang@hcl.com>