Specify what to match on for Forwarding. This is a first step towards automatically adapting the match depending on what the switch requires for the flow to be in hardware.

Fixed action serializer for actions that do not have values e.g. strip_vlan and such. Added in a 'n/a' value for the JSON key's value. Perhaps we should rethink how a list of actions is serialized to avoid the 'n/a' all together.

Fixed issue where multipart flow replies would only show the last part/reply and not print the others. Only a REST API issue. This will also fix the Web UI where people reported not seeing all their flows.

Modified slightly to track the specific initial state. If we were slave first, then we won't clear the flow table if we go to master for the first time.

Added ability to customize start/role-transition behavior at the request of an individual a while ago. Using floodlightdefault.properties, you can specify if you want the controller to clear the flow tables of switches (1) upon inital handshake and transition to master or (2) upon every transition to master. Or you can not ever clear the tables if you want.

Added ability to switch OF roles per DPID or for all switches at once. This includes a REST interface and integration with the handshake and channel handlers in order to keep the switch in the correct state. An IOFConnectionListener also has the ability to listen for messages that are written now instead of read from the wire. This allows modules to write role requests, have the underlying handshake/channel substrate tap in and listen to the request, and automatically update the role of the switch when the response comes back.

Fixed OFChannelHandler to always allow echo requests and replies no matter the state in the handshake process. This issue was discovered during an unsuccessful handshake with an HP 5900af, which for whatever reason wants to handshake as follows: ... (1) FL features request, (2) HP echo request, (3) FL echo reply, (4) HP features reply. There are still issues with this switch after the handshake completes, but an echo should be honored in any state.

Fixed fetching of ingress port to work with OF1.0 and 1.2+. This is the final commit for the DHCP server. Will add some unit tests in a later and near-future commit as practice/exercise for me =)

Initial DHCP server code commit. This was a test of how easy is is to port v0.91- code to v1.0. I have not tested this module yet. TODO: Write unit tests for the dhcpserver module.

Reapplied the LinkDiscoveryManager's patch from earlier (issue #471). Updated Ethernet.java to use EthType instead of short for EthTypes. In conjunction, bitmasked the LSBs of the ethertype as it is casted from short to int. This preserves the MSB (sign) as a data bit --> unsigned integer.

In line 329 of StaticFlowEntryPusherResource.java, the word "flow mod"
is spelled as "flod mod".
Just encountered that when using the rest api, so I fixed it.

Forwarding's doDropFlow() changed to not base Match on OFPacketIn, which it shouldn't in the first place. We need to first deserialize the packet and use its header attributes. TODO: How much is too much to match on? Sometimes we want a very specific match by default. Other times, we don't. The former creates a larger flow table, while the latter might inadvertently match packets that we don't intend to. Perhaps this should be a user-configurable option? Such as net.floodlightcontroller.Forwarding.matchLayer=L2, L3, or L4?

Undo LinkDiscoveryManager 'bug' fix from a while back. We'll test to see if it's what's causing some issues with invalid ethertypes. Also, includes a patch for OFChannelHandler that will accept the smallest of the OpenFlow versions during the initial handshake.

Updated Restlet from 2.2M3 to 2.3.1. Included with the update is full SSL support for the REST API, either all-access+encryption or restricted-access+encryption (truststore verified clients only). Additional options have been added to floodlightdefault.properties in order to turn HTTPS on/off. HTTPS ccan be used in conjunction with vanilla HTTP as well -- each must reside on a different TCP port though. Also, this time, I remembered to update and test the build.xml file to remove the old and include the new Restlet jars.

Bug fix WRT closed issue #471. ethertype range check for valid LLC lengths expanded to include 17-1535 inclusive and <17 as invalid.

SSL support for southbound OpenFlow sockets. Enable/disable SSL from floodlightdefault.properties (YES=enabled, NO=disabled to the useSsl variable), set the keystore, and password. By default, SSL is disabled. Keystore needs to be configured with a public/private/cert for FL and with the certificates for each OVS ahead of time. Each key should have the same password as the keystore itself.

Fixed a bug that caused flows without instructions/actions (i.e. drop flows) to throw an exception during serialization. This addresses an issue common to the SFP and the switch flow stats reply serialization (since they both rely on the same core.web serializer).

Fixed a bug where table ID, group ID, and cookie mask were attempted to be serialized for an OF1.0 switch. Should probably refactor this to make the code less repetetive now.

Fixed a tough bug in MatchUtils. OFPorts are integers in OF1.3, but those who specify local, controller, etc. as 65534, etc. will not work with OF1.3. This is a temporary fix and should be fixed properly to allow the full range of ports.

Let's try this again. Fixed the two SFP bugs (Maxence's bug and another that prevented adding drop flows with null or empty-string actions). Also, replaced netty with a newer version in preparation for SSL support.

Revert "Fixed a couple bugs in SFP. (1) Always check the switch we're pushing a new flow to to see if an old flow needs to be removed from another switch (Maxence's bug). (2) Check for empty string in action lists and only print/add actions if they are specified. This allows an empty string or null to mean 'drop.'"

This reverts commit e2458ef3.

Fixed a couple bugs in SFP. (1) Always check the switch we're pushing a new flow to to see if an old flow needs to be removed from another switch (Maxence's bug). (2) Check for empty string in action lists and only print/add actions if they are specified. This allows an empty string or null to mean 'drop.'

Don't use the Match from the Firewall in Forwarding if a rule allowed the packet. If a packet has been allowed, that means the Forwarding module can send it where it needs to go matching on its header files (which have already been given the OK by the Firewall). If a Firewall rule is very general, e.g. allow all packets through switch 1, then the first packet that traverses the switch will cause Forwarding to insert a general from from port A to port B with no specific hheader field matches (since they weren't specified in the Firewall rule).

Added a patch to support HP ProCurve switches and others that do not support the OFPT_BARRIER_REQUEST message. Updated SFP unit test to include the SEND_FLOW_REM flag in the test flows (forgot from last commit).

SFP supports hard/idle timouts w/drawback of needlessly sending flow_del to switch (which gets silently discarded). This is a good first step but needs to be followed up with a solution to the redundant flow_del msg. Also, patched MatchSerializer to display masks if present and fixed a bug in OFInstructionListSerializer for metadata and goto_table instructions. Added masking support to all SFP match fields. OVS supports some masked fields that are not defined in the OF1.3 spec. Oddly enough, OVS returns a bad-match error message even though the flow gets inserted. I suppose that message is due to the OF protocol not supporting it. OVS should squelch it if it supports it though. In any case, Floodlight is equipped to support all present MatchFields as masked MatchFields in the future. To specify a mask to any field (beside MPLS BOS, which is a boolean), use a CIDR-style '/' after the value and before the mask. All values and masks support decimal or hexadecimal (with leading '0x'), with the exception of all values that are MacAddress or IPAddress types -- they only support hex-string or dotted-decimal, respectively. Lastly, an English typo has been fixed in IStaticFlowEntryPusherService.

Removed some more unnecessary comments. Don't know how hard_timeout and idle_timeout got reverted in the SFP...