Completely disable use of bind() with pseudo-random port selection on Windows.

On Windows, completely disable use of bind() with pseudo-random port selection.

I'll land this on the tip-of-tree, and then revert it.  We'll then apply
it to the dev branch so that there is no chance of accidentally getting
security dialogs on Windows.

This should reduce the bug to P2, and allow a more complete resolution
of the bug.

r=rch
BUG=329255
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243858

[jar (doing other things), 2014-01-08 17:15:00]

Ryan/WanTeh: Could either of you review this.  It is minimal, and this is a
stop-gap fix, which will avoid using our RANDOMized bind() on Windows. 

I'll land this on the TOT, so that it can be merged into the Dev branch.  I'll
then revert on TOT and work on a cleaner solution that will only disable this
for windows on Beta and Stable.  Finally, we'll work to set a firewall rule to
avoid popup in most installations, and only call our bind() on windows when we
have the protection of that rule.

[wtc, 2014-01-08 17:40:13]

Patch set 3 LGTM.

[commit-bot: I haz the power, 2014-01-08 17:42:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jar@chromium.org/127463003/90001

[Ryan Hamilton, 2014-01-08 17:53:41]

https://codereview.chromium.org/127463003/diff/90001/net/quic/quic_stream_fac...
File net/quic/quic_stream_factory.cc (right):

https://codereview.chromium.org/127463003/diff/90001/net/quic/quic_stream_fac...
net/quic/quic_stream_factory.cc:517: bind_type = DatagramSocket::DEFAULT_BIND;
This change *should* break the test which verifies that the ephemeral port
selection is deterministic.  Can you figure out what is going on?

[commit-bot: I haz the power, 2014-01-08 19:43:51]

Retried try job too often on win_rel for step(s) net_unittests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[jar (doing other things), 2014-01-08 22:17:11]

PTAL  (neutered Window's testing of our use of bind())

[Ryan Hamilton, 2014-01-08 22:21:52]

lgtm

https://codereview.chromium.org/127463003/diff/380001/net/quic/quic_stream_fa...
File net/quic/quic_stream_factory.cc (right):

https://codereview.chromium.org/127463003/diff/380001/net/quic/quic_stream_fa...
net/quic/quic_stream_factory.cc:517: bind_type = DatagramSocket::DEFAULT_BIND;
nit: i wonder if it would more symmetrical to do:

#if defined(OS_WIN)
  // TODO(jar)bug=329255 Provide better implementation to avoid pop-up warning.
  DatagramSocket::BindType bind_type = DatagramSocket::DEFAULT_BIND;
#else
  DatagramSocket::BindType bind_type = DatagramSocket::RANDOM_BIND;
#endif

Clearly this is not terribly important, but doing an assignment in the ifdef
just looked a bit odd.

[wtc, 2014-01-08 22:36:05]

Patch set 4 LGTM.

https://chromiumcodereview.appspot.com/127463003/diff/380001/net/quic/quic_st...
File net/quic/quic_stream_factory.cc (right):

https://chromiumcodereview.appspot.com/127463003/diff/380001/net/quic/quic_st...
net/quic/quic_stream_factory.cc:534: #endif

I would just do

// TODO(jar)bug=329255 Provide better implementation to avoid pop-up warning.
#if !defined(OS_WIN)
  UMA_HISTOGRAM_COUNTS("Net.QuicEphemeralPortsSuggested",
                       port_suggester->call_count());
  DCHECK_LE(1u, port_suggester->call_count());
#endif

https://chromiumcodereview.appspot.com/127463003/diff/380001/net/quic/quic_st...
File net/quic/quic_stream_factory_test.cc (right):

https://chromiumcodereview.appspot.com/127463003/diff/380001/net/quic/quic_st...
net/quic/quic_stream_factory_test.cc:581: return;

Our convention is to name the test name as MAYBE_CreateConsistentEphemeralPort.

Search for "MAYBE_" in src/net for an example.

Doing it this way may generate a compiler warning about unreachable code on
Windows after the return statement.

[commit-bot: I haz the power, 2014-01-08 22:38:34]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jar@chromium.org/127463003/380001

[commit-bot: I haz the power, 2014-01-09 12:11:19]

Change committed as 243858

[jar (doing other things), 2014-01-09 20:54:41]

A revert of this CL has been created in
https://codereview.chromium.org/132603002/ by jar@chromium.org.

The reason for reverting is: Reverting on Canary to maintain port selection
there.  Will land on Dev branch..