Note that we are changing the interface of batchStartup/Shutdown
slightly so that the method can choose not to start some of the
switches. We might wish to refine this a bit...

This should rarely happen - in the usual case, either the
links will be shut down by Mininet.stop(), or the interfaces
will be deleted by node.stop( deleteIntfs=True ), or the
links or interfaces will be explicitly deleted or stopped
using the low-level API.

Cases that are relying on links being automatically deleted
in cleanup() will potentially find that they are now no longer
deleted, but these cases should be rare.

We still need to set it in batchShutdown()

This implements batch startup for OVS switches.

In the future, we would like to support higher bandwidth
limits for TCIntfs, but we'll probably need to adjust some
of the parameters dynamically.

In my opinion, we really shouldn't even get POLLHUP, but
we do. In case we get anything else odd, quite polling on
that fd.

It's tricky to get this right, but basically we want to read
if there is something to read; if not, we want to check for
EOF.

Without this, we end up retrying until we fail, instead
of returning immediately on success!

We should think a bit about the semantics that we want here.
The comments say "stop and clean up link" so perhaps that's
what we want. However, we could also imagine stop stopping
forwarding on the link (and possibly allowing restarts).

We warn on exited node.cmd() because we terminate the controller
before stopping/deleting the links. This makes sense to avoid a
storm of link/port down events, but since the controller's
shell has exited we cannot call link.stop() on any of its
links. We may want to simply stop the controller and not
terminate it, but at least it doesn't hang for now.

This allows --switch ovsbr,stp=True to work correctly

With newer versions of OVS, this allows us to set up a switch
with a single OVS command (if reconnectms is zero.) If reconnectms
is specified, then it slows things down slightly (but not much.)

Eventually we should implement true batch shutdown.
In the mean time, we just ignore it. Note there's no good
way that I know of for a subclass to remove a superclass
method, so we changed the protocol a bit to require a return
value of True.

Note that this also changes the way that links are deleted;
the reason is that the batch shutdown doesn't currently delete
the links, but OVSSwitch.stop() does. We may wish to revisit
this in the future.

It makes sense to follow Open vSwitch's lead here.
OVS 2.3 enables OpenFlow 1.0 through 1.3 by default.
OVS 2.0 has incomplete support for 1.3, but you can enable
it with protocols=OpenFlow13

For OVS versions 1.9 or earlier, this setting is ignored.