diff options
| author | rtenneti <rtenneti@chromium.org> | 2015-09-13 21:43:11 -0700 |
|---|---|---|
| committer | Commit bot <commit-bot@chromium.org> | 2015-09-14 04:47:33 +0000 |
| commit | 493d90ef42c523aa443191dbe7c6fab35be91abf (patch) | |
| tree | bf1ba7220cd4bea430f599c1e9cbeffea864a8b4 /remoting/protocol | |
| parent | 9418953296227e8a861bf8e3ea8f86a07904e273 (diff) | |
| download | chromium_src-493d90ef42c523aa443191dbe7c6fab35be91abf.zip chromium_src-493d90ef42c523aa443191dbe7c6fab35be91abf.tar.gz chromium_src-493d90ef42c523aa443191dbe7c6fab35be91abf.tar.bz2 | |
Landing Recent QUIC changes until 8/28/2015 18:03 UTC.
relnote: Refactor QuicAckFrame::missing_packets to support a change to a
more efficient data structure. No functional change.
Merge internal change: 101778041
https://codereview.chromium.org/1330973002/
relnote: n/a (changing DLOGs). Tidy up the DLOG messages on receipt of
various frame types. I was confused for a while by the connection close
message (my test server was *receiving* it, rather than closing the
connection itself), hence the change to that message. Other edits bring
consistency.
Merge internal change: 101773586
https://codereview.chromium.org/1304373010/
relnote: QUIC bugfix (and test) for setting smaller receive window size
under flow-control auto-tuning. Behavior gated by finch tag AFCW,
which is currently enabled in canary and dev.
Prior to this fix, streams established before the crypto handshake,
i.e. headers and crypto, would have their flow control window size
become out of sync between client and server, when tags to set
smaller initial window sizes were used (kIFW*). It is believed
this was causing some flow control deadlocks in the auto-tuning trial.
This change also corrects an oversight, adding logic so that kIFW*
reduce the initial session flow control window size as well.
Merge internal change: 101695854
https://codereview.chromium.org/1327903004/
relnote: Add num_connections_ in quic_time_wait_list_manager to record
number of connection IDs in time wait list. (No functional change)
Merge internal change: 101619602
https://codereview.chromium.org/1310753008/
relnote: Allow individual QUIC writers to limit the maximum packet size.
Protected by FLAGS_quic_limit_mtu_by_writer, enabled by default.
Since this CL, all QuicConnections are required to have a valid peer
address when they are created.
FIXED=23280955
Merge internal change: 101617105
https://codereview.chromium.org/1320303009/
R=rch@chromium.org, sergeyu@chromium.org
Review URL: https://codereview.chromium.org/1331053003
Cr-Commit-Position: refs/heads/master@{#348555}
Diffstat (limited to 'remoting/protocol')
| -rw-r--r-- | remoting/protocol/quic_channel_factory.cc | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/remoting/protocol/quic_channel_factory.cc b/remoting/protocol/quic_channel_factory.cc index 0a49546..26ff3c8 100644 --- a/remoting/protocol/quic_channel_factory.cc +++ b/remoting/protocol/quic_channel_factory.cc @@ -23,6 +23,7 @@ #include "net/quic/quic_clock.h" #include "net/quic/quic_connection_helper.h" #include "net/quic/quic_default_packet_writer.h" +#include "net/quic/quic_protocol.h" #include "net/socket/stream_socket.h" #include "remoting/base/constants.h" #include "remoting/protocol/datagram_channel_factory.h" @@ -77,6 +78,10 @@ class P2PQuicPacketWriter : public net::QuicPacketWriter { } bool IsWriteBlocked() const override { return write_blocked_; } void SetWritable() override { write_blocked_ = false; } + net::QuicByteCount GetMaxPacketSize(const net::IPEndPoint& peer_address) const + override { + return net::kMaxPacketSize; + } private: void OnSendComplete(int result){ |