Land Recent QUIC Changes

Land Recent QUIC Changes

Fix to QUIC's RttStats ExpireSmoothedMetrics where the mean deviation
may increase if the rtt decreased.

Merge internal change: 71665066
https://codereview.chromium.org/406403002/


Allow client to request Reno congestion control.

Disabled BBR tests (will enable after BBR code is checked in).

Merge internal change: 71621257
https://codereview.chromium.org/412753002/


QUIC - Minor cleanup change to keep internal source tree and chromium
codebase in sync.

+ Use DECHEK_EQ instead of DCHECK.

Merge internal change: 71613624
https://codereview.chromium.org/412733002/


Allow QuicConnection to own its QuicPacketWriter

Added a new "bool owns_writer" parameter to the QuicConnection
constructor that specifies whether the QuicConnection has ownership.
Also cleaned up the ownership semantics in a number of tests --
previously, the writer was not actually outliving the connection.

Merge internal change: 71590341
https://codereview.chromium.org/413573003/


Added QuicTime::Infinite() to create infinite time required by the
following internal change.

Change QUIC"s BBR sender to only fall out of slow start when there is a
packet loss if the residual is less than 0.25.

Merge internal change: 71436337
https://codereview.chromium.org/410743003/


Ensure the UnackedPacketMap only keeps one previous transmission less
than largest_observed.

Also improves the tightness of QuicUnackedPacketMapTest.

Merge internal change: 71426216
https://codereview.chromium.org/411723003/


Parameterize the QUIC end-to-end tests to run with FEC both enabled and
disabled.

Found bug that was fixed in internal change: 71390633
(https://codereview.chromium.org/414523003/)

Merge internal change: 71393479
https://codereview.chromium.org/412543005/


Fix a bug in the QuicReceivedPacketManager where a packet which had
previously been revied was not removed from the revived packet set when
it was eventually received out of order.

Merge internal change: 71390633
https://codereview.chromium.org/414523003/


QuicConnection::ProcessValidatedPacket doesn't need to be public because
it is only called by QuicConnection::OnPacketHeader.

Merge internal change: 71371063
https://codereview.chromium.org/402333003/


Add bandwidth_estimate and server_namespace fields to the source-address
token proto. Not used anywhere, this CL is really just for discussion
about which fields (and fieldnames) we should use.

Merge internal change: 71360060
https://codereview.chromium.org/404343003/

R=rch@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=287052

[Ryan Hamilton, 2014-07-23 03:40:24]

lgtm

Might be worth holding off landing this until we get the perf issue sorted out.
What do you think?

[ramant (doing other things), 2014-07-23 04:25:53]

On 2014/07/23 03:40:24, Ryan Hamilton wrote:
> lgtm
> 
> Might be worth holding off landing this until we get the perf issue sorted
out.
> What do you think?

+1. Wont submit. Will wait for green light from you.

Thanks very much Ryan.

[wtc, 2014-07-23 23:31:42]

Patch set 1 LGTM.

https://codereview.chromium.org/411823002/diff/1/net/quic/congestion_control/...
File net/quic/congestion_control/rtt_stats.cc (right):

https://codereview.chromium.org/411823002/diff/1/net/quic/congestion_control/...
net/quic/congestion_control/rtt_stats.cc:48:
std::abs(smoothed_rtt_.Subtract(latest_rtt_).ToMicroseconds())));

Why do we need to convert to microseconds and then convert back? So we can take
the absolute value?

[ramant (doing other things), 2014-07-23 23:56:47]

Fixed wtc's comments for CL: https://codereview.chromium.org/404343003/

https://codereview.chromium.org/411823002/diff/1/net/quic/congestion_control/...
File net/quic/congestion_control/rtt_stats.cc (right):

https://codereview.chromium.org/411823002/diff/1/net/quic/congestion_control/...
net/quic/congestion_control/rtt_stats.cc:48:
std::abs(smoothed_rtt_.Subtract(latest_rtt_).ToMicroseconds())));
On 2014/07/23 23:31:42, wtc wrote:
> 
> Why do we need to convert to microseconds and then convert back? So we can
take
> the absolute value?

Seems like that.

Ian Swett: wdyt?

[ramant (doing other things), 2014-07-25 23:10:05]

The CQ bit was checked by rtenneti@chromium.org

[commit-bot: I haz the power, 2014-07-25 23:12:09]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/rtenneti@chromium.org/411823002/60001

[ramant (doing other things), 2014-07-25 23:12:29]

On 2014/07/23 04:25:53, ramant wrote:
> On 2014/07/23 03:40:24, Ryan Hamilton wrote:
> > lgtm
> > 
> > Might be worth holding off landing this until we get the perf issue sorted
> out.
> > What do you think?
> 
> +1. Wont submit. Will wait for green light from you.
> 
> Thanks very much Ryan.

Wan-Teh gave the green light. Landing it.

[ramant (doing other things), 2014-07-26 00:07:45]

The CQ bit was unchecked by rtenneti@chromium.org

[ramant (doing other things), 2014-07-26 00:08:16]

On 2014/07/25 23:12:29, ramant wrote:
> On 2014/07/23 04:25:53, ramant wrote:
> > On 2014/07/23 03:40:24, Ryan Hamilton wrote:
> > > lgtm
> > > 
> > > Might be worth holding off landing this until we get the perf issue sorted
> > out.
> > > What do you think?
> > 
> > +1. Wont submit. Will wait for green light from you.
> > 
> > Thanks very much Ryan.
> 
> Wan-Teh gave the green light. Landing it.

Will wait until next week. Thanks Ryan/Wan-Teh.

[ramant (doing other things), 2014-07-31 21:52:15]

The CQ bit was checked by rtenneti@chromium.org

[commit-bot: I haz the power, 2014-07-31 21:54:11]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/rtenneti@chromium.org/411823002/100001

[commit-bot: I haz the power, 2014-08-01 10:44:30]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  win_chromium_x64_rel on tryserver.chromium.win
(http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[commit-bot: I haz the power, 2014-08-01 18:42:45]

Change committed as 287052