Don't unmap ports in host_driven/test_case.py.

Don't unmap ports in host_driven/test_case.py.

The sync tests failure seems to have to do with port unmapping. Not unmapping
ports let most of the sync tests pass (only 2 failures remain for an unrelated
reason apparently).
Note that port unmapping was added to avoid exceeding the file descriptor limit
in net_unittests since each test server was listening on a different port. The
sync tests use a server always listening on a same port so unmapping the port
is not strictly needed.

BUG=313809
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=239467

[Philippe, 2013-12-05 13:51:17]

Hi Marcus,

I have this partial workaround for the sync tests downstream (that seem to be
still failing :/). I'm not proud of it but it seems to improve the situation
somehow (although 2 tests seem to be still failing).

[bulach, 2013-12-06 10:53:19]

lgtm but...ouch... :(

what is the issue with unmapping? why is it failing?
this seems that the tests are abusing previously mapped ports rather than map
their own... if the ports are in a bad state due to the _previous_ test, this
will potentially make more tests fail, no?

we should really consider having a forwarder process per device+port, so that
all this cleanup will be much simpler..

[Philippe, 2013-12-09 09:21:26]

I don't fully understand what's happening here to be honest but the problem
seems to be related to the unmapping of the port for the sync server in
particular. The unmapping for the HTTP server on port 8000 still happens for
each unit test through base_test_runner.py though. Keeping a port mapped should
not be too bad unless one of the layer gets out of sync and loses the knowledge
that it's mapped (which should not happen here). I agree that a separate process
for each device/port would be good to improve scalability although I'm not sure
it would solve this particular problem :/

[commit-bot: I haz the power, 2013-12-09 09:21:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/pliard@chromium.org/106533002/1

[commit-bot: I haz the power, 2013-12-09 11:13:28]

Change committed as 239467