Land Recent QUIC changes.

Land Recent QUIC changes.

Merging changes from chromium CL - 16203007

Merge internal change: 48035020

Minor changes from review comments from CL 48035020.

Fix http://crbug.com/248737 where QuicSession::ConnectionClose
would DCHECK when attempting to close a stream that did not close
via ConnectionClose.  Add regression test.

Merge internal change: 47988326

Added a helper method SetSelfAddress to set the real address. 


Merge internal change: 47983053

Allow packets to be sent by the client when encryption is first established
without waiting for the handshake to be confirmed.  Simply doing this made
the LargePost test flaky.  To deal with *that* problem, I made two more
changes:

  1) When setting up an RTO for an ENCRYPTION_NONE packet, do not
     exponentially back off.
  2) When an endpoint receives encrypted packets that can not be
     decrypted buffer some of them and attempt to re-process them
     when the endpoint moves to the next encryption level.

Merge internal change: 47934275

Fix grammatical errors in comments.

Merge internal change: 47877973

QUIC: allow replay protection to be configured.

Bandaid nodes can't run a shared strike-register (at least,
not without a lot of trouble) so would need different orbits
for each machine.

However, we'll be sending HTTP requests on the QUIC connection anyway - and
will be falling back to HTTP in the event of a failure. So there's no actual
security. Rather than burn round trips having the clients fetch new server
configs each time they talk to a new Bandaid machine, just disable replay
protection and give them all the same server config.

This change is the QUIC part of disabling replay protection. The GFE config
still remains to be done, as does the distribution of server configs to the
Bandaid machines.

Merge internal change: 47857583

Removing an unnecessary API from the quic stream sequencer, cleaning up
an obsolete GFE TODO.  We propagate stream RSTs immediately
(via PropogateClose gfe-side) so no longer need to pull them from the
sequencer.

Merge internal change: 47848450

Minor cleanup change for SetNonce in StrikeRegisterTest per wtc's comments
in CL: https://codereview.chromium.org/15937012/

R=rch@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=207016

[ramant (doing other things), 2013-06-17 18:42:41]

Hi Ryan,
  This is a merge of all CL's (7 CLs of yours and 1 CL of wtc) you have
approved.

rch approved CLs:
17106005
17043010
17114004
16951020
17113003
17101003
16888009

wtc's approved CL's 17040003

[Ryan Hamilton, 2013-06-17 22:40:55]

lgtm

[commit-bot: I haz the power, 2013-06-17 22:55:36]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtenneti@chromium.org/17302002/1

[commit-bot: I haz the power, 2013-06-18 13:53:00]

Commit queue rejected this change because the description was changed
between the time the change entered the commit queue and the time it
was ready to commit. You can safely check the commit box again.

[commit-bot: I haz the power, 2013-06-18 16:38:50]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtenneti@chromium.org/17302002/1

[commit-bot: I haz the power, 2013-06-18 16:39:29]

Change committed as 207016