Marten Seemann
4076ab587e
add a string representation for the Long Header packet types
2017-11-22 15:57:22 -08:00
Jan Rüth
281fec1b2d
Make error chan not block for a single write, allowing ParseHandshakeMessage goroutine to gracefully exit
2017-11-22 09:54:03 +01:00
Marten Seemann
683f244054
rename the Cleartext Long Header type to Handshake
...
This was recently changed in the draft.
2017-11-13 10:55:33 +08:00
Marten Seemann
7c3d6abb4b
fix parsing of the Header type byte
...
In order to determine if a packet is a Version Negotiation Packet, it is
not sufficient to just look at bit 0x1. Other packet types also have
that bit set, e.g. the Retry packet (packet type 0x3). Instead, we have
to look at the last 3 bits. This fix will work as long as IETF QUIC
doesn't define more than 8 long header types.
2017-11-13 10:23:17 +08:00
Marten Seemann
234d5aab36
echo the offered version in IETF Version Negotiation Packets
2017-11-12 14:37:43 +08:00
Marten Seemann
11364cbd81
fix encoding of the idle timeout for the TLS transport parameters
2017-11-11 19:48:41 +08:00
Marten Seemann
5d4174b2f0
parse the packet header before getting the session from the session map
2017-11-10 08:48:45 +07:00
Marten Seemann
bc1ef31531
accept Public Resets without the RSEQ tag
...
The RSEQ tag is mandatory according to the gQUIC spec, however, Google
doesn't include that tag in its Public Resets.
2017-11-08 11:07:53 +07:00
Marten Seemann
095c29dc2c
drop support for QUIC 37 and 38
2017-11-07 00:23:53 +07:00
Marten Seemann
9f1eb62860
implement the IETF draft BLOCKED and STREAM_BLOCKED frames
2017-11-04 15:22:14 +07:00
Marten Seemann
0f1f1c8d41
implement the MAX_DATA and MAX_STREAM_DATA frames
...
For gQUIC WINDOW_UPDATEs are converted to MAX_DATA and MAX_STREAM_DATA,
respectively.
2017-11-04 15:22:14 +07:00
Marten Seemann
1f644debd4
set if a stream contributes to connection flow control
...
This depends on the version. In gQUIC, stream 1 and 3 don't contribute,
in IETF QUIC only stream 0 doesn't contribute.
2017-11-04 15:22:14 +07:00
Marten Seemann
f662822486
use stream 0 for the crypto stream when using TLS
2017-11-04 15:22:07 +07:00
Marten Seemann
a1a4f35096
send initial version in the CHLO, not the current version
...
This commit fixes two bugs:
1. In the CHLO, we need to send the initial QUIC version. It will differ
from the current version if version negotiation was performed.
2. The session setup was using the wrong version (current version, and
not the initial version), such that we would have sent the wrong version
in the TLS handshake as well.
2017-11-02 21:51:51 +07:00
Marten Seemann
96f65f922f
Merge pull request #927 from lucas-clemente/fix-925
...
fix client handshake failure when the server supports unknown versions
2017-11-02 21:34:04 +07:00
Marten Seemann
b40d249953
reject packets that use the wrong or unknown packet types
2017-11-02 20:47:02 +07:00
Marten Seemann
3e6f66da79
rename the packet types according to recent draft changes
2017-11-02 18:31:04 +07:00
Marten Seemann
49e305f97f
Merge pull request #915 from lucas-clemente/fix-914
...
only use little endian byte order for gQUIC 37 and 38
2017-11-02 17:59:55 +07:00
Marten Seemann
43c23ff5fa
fix client handshake failure when the server supports unknown versions
2017-11-02 17:41:06 +07:00
Marten Seemann
aba1dd13ba
only use little endian byte order for gQUIC 37 and 38
...
That way, when adding new non-gQUIC versions, they will use big endian.
2017-11-02 16:46:22 +07:00
Marten Seemann
df13a0a970
implement the TLS Cookie extension
2017-11-02 16:45:43 +07:00
Marten Seemann
cc5f978996
Merge pull request #912 from lucas-clemente/fix-911
...
fix gQUIC version numbers
2017-10-30 19:46:36 +07:00
Marten Seemann
b0f34e776e
send IETF draft style version negotiation packets
...
Send a gQUIC Version Negotiation Packet, if the client packet has a
Public Header. If the client has an IETF draft style header, send an
IETF draft style Version Negotiation Packet.
2017-10-30 18:23:12 +07:00
Marten Seemann
f6e44e3e8e
fix gQUIC version numbers
2017-10-30 11:51:26 +07:00
Marten Seemann
3f62ea8673
set the Long Header packet type based on the state of the handshake
2017-10-29 14:08:08 +07:00
Marten Seemann
a65929f6cf
add logging for the packet header
2017-10-27 22:56:41 +07:00
Marten Seemann
3e39991e1e
use the non-blocking of mint to cycle through the handshake
2017-10-27 22:54:40 +07:00
Marten Seemann
fcc380187a
Merge pull request #899 from lucas-clemente/fix-898
...
encrypt unprotected packets with a connection-dependent key
2017-10-27 22:52:04 +07:00
Marten Seemann
4f86995e48
encrypt unprotected packets with a connection-dependent key
2017-10-27 17:35:08 +07:00
Marten Seemann
ddea1d83ab
use gomock to mock crypto.AEAD in the crypto setup tests
2017-10-27 17:35:08 +07:00
Marten Seemann
d98a11bb35
save version numbers such that can be written in big endian
...
This makes the version number representation consistent with the IETF
draft.
2017-10-27 17:24:24 +07:00
Marten Seemann
e81795e49b
Merge pull request #895 from lucas-clemente/new-header
...
implement the IETF draft Header
2017-10-27 17:18:15 +07:00
Marten Seemann
4556ad01e5
use the new packet header for connections using TLS
2017-10-27 16:54:47 +07:00
Marten Seemann
6c394acde7
never use a 6 byte packet number
...
According to the IETF draft, the only packet number lengths are 1, 2 and
4 bytes. With the given formula for the packet number derivation, we
would have only sent a 6 byte packet number if the difference between
the largest unacked and the current packet number exceeded 2^31, so this
would never have happened anyway.
2017-10-27 16:51:59 +07:00
Marten Seemann
bbe9affa27
reject version negotiation packets with an empty version list
2017-10-27 16:51:59 +07:00
Marten Seemann
7cb8e87580
implement parsing of IETF Version Negotiation Packets
2017-10-27 16:51:57 +07:00
Marten Seemann
fad279cb75
implement parsing and writing of the IETF QUIC Header
2017-10-27 16:51:04 +07:00
Marten Seemann
d7a09e7667
reject inconsistent final stream offsets
2017-10-27 16:39:41 +07:00
Marten Seemann
636bf4578c
error when receiving stream data with a larger offset than the final offset
...
This could happen when a peer send a STREAM frame with the Fin bit set,
and then sends STREAM frames with a higher offset.
2017-10-25 10:43:29 +07:00
Marten Seemann
ac82aeda8b
treat the connection ID as a big endian value
2017-10-25 10:09:15 +07:00
Marten Seemann
9825ddb43a
simplify the interaction with mint
2017-10-25 09:13:47 +07:00
Marten Seemann
282b423f7d
pass the crypto stream to the crypto setup constructor
...
The crypto stream is opened during the session setup. Passing it to the
crypto setup directly helps simplify the constructor.
2017-10-25 09:13:47 +07:00
Marten Seemann
a88da29433
Merge pull request #881 from lucas-clemente/improve-flow-controller
...
remove the flow control manager
2017-10-25 09:10:39 +07:00
Marten Seemann
e49ef8a90c
Merge pull request #889 from lucas-clemente/fix-connection-close-parsing
...
fix parsing of long reason phrases in CONNECTION_CLOSE frames
2017-10-25 09:10:16 +07:00
jbenoist
12d4e031bf
fix 0-RTT by moving cookie generation to server config
2017-10-24 13:27:50 +07:00
Marten Seemann
5556edbcad
fix parsing of long reason phrases in CONNECTION_CLOSE frames
2017-10-23 17:23:13 +07:00
Marten Seemann
268c3859fc
remove the flow control manager
2017-10-20 21:34:09 +07:00
Marten Seemann
6dc453caa3
use mockgen reflect mode
...
This requires a dirty hack to get mockgen to generate mocks for internal
packages.
2017-10-20 21:12:19 +07:00
Marten Seemann
e087ee7e9f
handle WINDOW_UPDATEs for streams and connection separately
...
In IETF QUIC, stream 0 is a valid stream ID, and is not used to encode
WINDOW_UPDATEs for the connection any more.
2017-10-20 21:12:19 +07:00
Marten Seemann
c53a83535e
split the flow controller in stream and connection flow controller
2017-10-20 21:12:19 +07:00
