fix-bug: in snmpit_hp.pm:getUsedOpenflowListenerPorts the port may be uninitialized if the listener string is empty. fix-bug: in snmpit.in:doEnableOpenflowListener Openflow listener output is printed out no matter whether the stacks support Openflow or not.

to keep in mind the possible change of the OF listener OID.

getUsedOpenflowListenerPorts in snmpit stack modules and devices are changed to return the ports rather than accept a hash reference. Add the empty OF functions to snmpit_nortel.pm. Print the OF listener string on each involved device. Fix a bug that snmpit.in doesnot print Done when OF operations are finished. The lengths of OF usage lines are reduced to fit the 80-ch width window.

fix bug: in snmpit_hp module, the SNMP session with MIB returns the OID as strings not numbers even for Openflow entry, however, the meaningful string OID cannot be used to get the next entry in MIB table! So getUsedOpenflowListenerPorts has to modify the OID of the current entry to be the numbered format.

Use a var instead of magic string in snmpit_hp to test OF listener var name. Add Openflow to OF related functions in snmpit HP module.

Add getUsedListenerPorts in HP snmpit module, it gives the used ports of existing listeners. This function is for automatic listener port assigning.

Openflow OIDs can be used in snmpit_hp now, so the functions can act correctly. Command line options configuration is changed to use only one vlanname for each OF command. The order of OF support checking and finding VLan is changed so that stack module will find vlan first, then check OF support. Thus there will not be an error msg for OF unsupported even if the vlan is not on the switch.

Openflow command line options are modified to have pid and eid so that stacks can be found according to experiment. snmpit.in has been modified to support multiple stacks.

lumping the VLAN number in with the OID, leaving the field we expect
to contain the VLAN number blank. I'm not sure why this happens, but
I've seen it before, and I think that simply dealing with it is going
to be more successful than finding the root cause.

Clean up a couple of minor inconsistancies in the modules concerning
printing ports not in the wires table and make sure utterances like
"wap snmpit -i Nortel10 4/4" work properly.

Better support for setting half- or full- duplex on ports.

still be considered experimental, but I beleive DETER has been running
it for some time now.

Note: I have not modified these files (including the copyright) in any
way: ie. Keith himself put the EMULAB-LGPL tag in the header of the
new file.

to the Cisco code too, but I want to do more investigation of them
before committing.

for snmpit.

Contributed by Keith Sklower at Berkeley.