Marten Seemann
595f6f814a
introduce a quic.Config.HandshakeIdleTimeout, remove HandshakeTimeout
2020-12-08 16:12:39 +07:00
Marten Seemann
3cd8b30ea6
Merge pull request #2856 from lucas-clemente/conn-id-update-handshake-completion
...
wait until the handshake is complete before updating the connection ID
2020-11-10 20:52:09 +07:00
Marten Seemann
e8a9888f25
Merge pull request #2857 from lucas-clemente/fix-spa-conn-id-handling
...
only check the SCID for Initial packets
2020-11-10 20:51:28 +07:00
Marten Seemann
0c2f562411
allow an amplification factor of 3.x
2020-11-06 14:16:11 +07:00
Marten Seemann
337e2171ef
only check the SCID for Initial packets
...
There's no risk that Handshake packets could be injected by an attacker.
2020-11-02 16:24:02 +07:00
Marten Seemann
80534c0944
wait until the handshake is complete before updating the connection ID
2020-11-02 16:21:38 +07:00
Marten Seemann
5c5e1024e7
send STREAMS_BLOCKED frame when MAX_STREAMS frame allows too few streams
2020-10-30 15:28:09 +07:00
Marten Seemann
8752576f26
run gofumpt, enable the gofumpt linter
2020-10-26 09:33:35 +07:00
Marten Seemann
145e7b10d0
Merge pull request #2811 from lucas-clemente/fix-first-key-update
...
allow the first key update immediately after handshake confirmation
2020-10-06 20:23:19 +07:00
Marten Seemann
55a07c34ee
add the exhaustive linter
2020-10-05 13:47:57 +07:00
Marten Seemann
bbd9fa4862
remove the EncryptionUnspecified enum value
2020-10-05 13:17:06 +07:00
Marten Seemann
b9090d71ae
rename cryptoSetup.DropHandshakeKeys() to SetHandshakeConfirmed()
2020-09-30 12:14:16 +07:00
Marten Seemann
b4636469fa
refactor confirmation of the handshake
2020-09-30 12:05:33 +07:00
Marten Seemann
ea3d32394d
read the ECN bits
2020-09-15 10:51:22 +07:00
Marten Seemann
fa4f0a9e7a
keep track of ECN counts on received packets
2020-09-14 17:01:34 +07:00
Marten Seemann
a528c4c4da
close session on errors unpacking errors other than decryption errors
2020-09-09 23:59:24 +07:00
Marten Seemann
e0824e3163
move the logging mock implementations to a separate package
2020-09-09 15:15:49 +07:00
Marten Seemann
556bf18dbf
inject a random source into the token protector
2020-08-23 17:06:17 +07:00
Marten Seemann
89417ab5ca
simplify the connection, rename it to sendConn
2020-08-09 19:24:04 +07:00
Marten Seemann
8db76ab9c2
only arm the application data PTO timer after the handshake is confirmed
2020-07-27 16:40:21 +07:00
Marten Seemann
0b7efe10d1
trace congestion state changes
2020-07-23 11:53:59 +07:00
Marten Seemann
0d4aa4b34f
introduce a function receivedPacket.Size()
2020-07-15 19:30:10 +07:00
Marten Seemann
a1bb39d6ab
introduce a protocol.StatelessResetToken
2020-07-10 19:55:20 +07:00
Marten Seemann
0ef1b2f92e
pass around the stateless reset token directly, not pointers to it
...
Benchmarks show that it's actually faster to make a copy of this 16 byte
array than passing around a pointer to it.
2020-07-10 19:44:57 +07:00
Marten Seemann
1f676c2a6c
use an errors.As comparison to detect stateless resets in the session
2020-07-10 19:43:54 +07:00
Marten Seemann
f93708aca8
introduce a logging.CloseReason to log the reason a session was closed
...
This also allows us to get rid of the ReceivedStatelessReset() method.
2020-07-10 19:23:32 +07:00
Marten Seemann
7d6ce4ea45
rename the logging.CloseReason to TimeoutReason
2020-07-10 19:17:21 +07:00
Marten Seemann
f16d474e7c
Merge pull request #2664 from lucas-clemente/move-versions-out-of-the-header
...
move the SupportedVersions slice out of the wire.Header
2020-07-10 18:24:46 +07:00
Marten Seemann
61d8e111b1
don't send write error in CONNECTION_CLOSE frames
2020-07-09 18:09:20 +07:00
Marten Seemann
993d71fd56
move the SupportedVersions slice out of the wire.Header
2020-07-09 12:05:33 +07:00
Marten Seemann
25db2166dd
enforce that a conn ID is not retired in a packet that uses that conn ID
2020-07-08 12:06:57 +07:00
Marten Seemann
ed120f670c
add logging aliases for frames, create a logging.STREAM and CRYPTO frame
2020-07-06 09:21:32 +07:00
Marten Seemann
865332015c
rename ResetStreamFrame.ByteOffset to FinalSize
2020-07-02 16:41:43 +07:00
Marten Seemann
fa07078d81
rename DataBlockedFrame.DataLimit to MaximumData
2020-07-02 16:41:43 +07:00
Marten Seemann
5466de6565
rename MaxStreamDataFrame.ByteOffset to MaximumStreamData
2020-07-02 16:41:43 +07:00
Marten Seemann
f4519cfd32
rename MaxDataFrame.ByteOffset to MaximumData
2020-07-02 16:41:43 +07:00
Marten Seemann
4121ea84f4
rename the logging.Tracer to ConnectionTracer
2020-07-01 16:57:01 +07:00
Marten Seemann
7f6eca84da
replace the logging.Export() method by a Close() method
2020-07-01 16:54:44 +07:00
Marten Seemann
7fdace490c
rename qlogger (member) variables to tracer
2020-07-01 16:53:45 +07:00
Marten Seemann
78ba619a0b
split the qlog package into a logging and a qlog package
2020-07-01 16:49:24 +07:00
Marten Seemann
06ad477b9b
handle Version Negotiation packets in the session
2020-07-01 14:19:45 +07:00
Marten Seemann
fe622dd780
use a token bucket pacing algorithm
2020-06-22 14:31:37 +07:00
Marten Seemann
4f19b15670
set and verify the retry_source_connection_id TP
2020-05-29 19:57:39 +07:00
Marten Seemann
fb2a5f413e
set and verify the initial_source_connection_id
2020-05-29 19:50:14 +07:00
Marten Seemann
b391cce35c
always send the original_destination_connection_id TP as a server
2020-05-29 19:50:14 +07:00
Marten Seemann
7e62a2fecc
rename original_connection_id to original_destination_connection_id TP
2020-05-29 19:50:14 +07:00
Marten Seemann
8d00ec135e
Merge pull request #2569 from lucas-clemente/drop-duplicate-packets
...
drop duplicate packets
2020-05-29 16:39:46 +07:00
Marten Seemann
60a918a108
limit available window to 3x of received bytes before address validation
2020-05-27 09:40:50 +07:00
Marten Seemann
98233f6743
add a way to pack coalesced packets smaller than the usual packet size
2020-05-27 09:33:34 +07:00
Marten Seemann
440ff107a3
drop duplicate packets
...
Duplicate detection uses the same data structure that is used to track
received packets to generate ACK frames. That means that after an old
ACK range has been pruned, a severly delayed packet might be incorrectly
detected as a duplicate.
As we wouldn't have acknowledged receipt of this packet, this situation
would have resulted in a retransmission by the peer anyway, so dropping
the packet won't cause a big regression.
2020-05-27 09:20:51 +07:00
