dtls_transport will when detecting a new fingerprint
(e.g by usage of pranswer) signal DtlsTransportState::kNew.
When this happen, the dtls crypto state is lost, and
sctp should reconnect, srtp does this automatically
in current code base.
The existing behavior in dcsctp is that it will detect
peer sending an init, and reconnect. But any messages sent
between the dtls restart and the message arriving from the
peer will be lost.
This patch changes so that this case is gracefully handled by
a) letting dcsctp_transport listen to dtls state
this is big part of patch and involves changing the type of
the underlying dtransport from rtc::PacketTransportInternal to cricket::DtlsTransportInternal. If requested, I can put this
into a separate patch...
b) if a dtls restart happens, delete and restart socket.
Testcase that fails before patch and works after is attached.
Bonus: And include-what-you-use on patch
Bug: b/375327137
Change-Id: Ib78488ae75fd8aeb50d121adf464a33dabbf95e2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/367202
Commit-Queue: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Victor Boivie <boivie@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Auto-Submit: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43546}
also fix a few TODOs
Bug: None
Change-Id: I2d287ed1a58f71ef32d5dc5624879ae8c63348b5
No-Try: True
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/370122
Reviewed-by: Fredrik Solenberg <solenberg@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43485}
The new type PriorityValue is a strong 16-bit integer matching RFC 8831
requirements that can be built from a Priority enum.
The value is now propagated and used by the SCTP transport, but enabling
the feature still requires a field trial for now.
Bug: webrtc:42225365
Change-Id: I56c9f48744c70999a8c2d01415a08a0b6761df4b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/357941
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42695}
If the field trial WebRTC-DataChannelMessageInterleaving is set, message
interleaving in SCTP (RFC8260) will be enabled in dcSCTP.
Bug: webrtc:41481008
Change-Id: I989b9ca554439ab0afd71f04d14a5cb5444b3361
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/354480
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42469}
This CL removes the send buffers (but not the receive buffer) from
SctpDataChannel and increases the send buffer in DcSctpSocket instead.
The reasons are:
1) Simplify the code. This additional buffering was strictly needed
before we migrated away from usrsctp, as that send buffer was very
limited in size (by design). But with the migration to dcSCTP, it's
no longer needed, so it just adds complexity.
2) Make `RTCDataChannel::bufferedAmount` correct. Before this CL, it
represented just the data buffered in SctpDataChannel, and not the
data accepted by the SCTP socket, but not yet put on the wire. This
makes it hard for clients to know when a message has ever been sent.
3) Better handle draining data on data channel close. While this is not
implemented in dcSCTP, having a single buffer makes this easier to
add.
While most of this CL is straightforward, the handling of bufferedAmount
in the signaling thread (in RTCDataChannel in Blink), is a bit special.
The number returned by `RTCDataChannel::bufferedAmount` is not what the
true value is inside the SCTP socket, but an eventual consistent view
of that value. When a message is sent, the value is incremented and:
- Before this change: When a message was put on the SCTP socket, the
view's value was decremented. Which made the view reflect what was
buffered outside the SCTP socket, and that buffering is now gone.
- After this change: SctpDataChannel will track what RTCDataChannel
will think it is, and provide updates to that number as we are
notified that it's reduced - by setting a "low threshold" callback
trigger.
A bonus with the new behavior is that it will be eventually consistent
and auto-heal also in error conditions - when messages are dropped due
to errors (bad input, bad state, etc). Previously, the bufferedAmount
value could drift away from the correct value on errors.
Note that a big chunk of unit tests were removed with this CL, as those
tested how the buffering behaved. Now, there is no buffering, so the
removed test cases represent a simpler interface.
This CL has been extensively tested with data channel benchmarks that
use the bufferedAmount thresholds (in Javascript).
Bug: chromium:40072842
Change-Id: I1a6a4af6b6e1116832f5028f989ce9f44683d229
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/343361
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41945}
This code was extracted to make the next following CL easier to review.
This CL simply exposes the getters, setters and callbacks to set the
buffered amount low threshold on a specific SCTP stream. It will be
used in a follow-up CL, but is just boilerplate.
Bug: chromium:40072842
Change-Id: Iccd72208b369ddc252cc5886f6446b9c2ceeb0b1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/343360
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41943}
This restricts the interface such that only a single onclose handler
can be set and that only one OnClose() notification will be fired.
That behavior is the same as how the previous sigslot was being
used, but the difference is that, in addition to removing sigslot,
this pattern is now more explicitly checked in the design.
Bug: webrtc:11943
Change-Id: I469c8cab3d62544988c8145b326af60b06b76d8e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/343340
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41920}
Before this change, calling buffered_amount only included what was
buffered on top of what was already buffered in the SCTP socket. With
the defaults, the SCTP socket can buffer up to 2MB of data (that is not
put on the wire) before the additional external bufferering in
SctpDataChannel will be used. The buffering that I am working on
removing completely.
Until it's removed completely, to avoid the issue reported in
crbug.com/41221056, include the bytes buffered in the SCTP socket to
what is returned when calling RTCDataChannel::buffered_amount.
This means that when this value is zero, it can be safe to know that all
bytes have been sent, but not necessarily acknowledged. And calling
close will not discard any messages.
This is a stopgap solution, but as functional as the proper solution
that removes all additional buffering. Follow-up CLs will merely improve
this solution.
Bug: chromium:41221056
Change-Id: I06edd52188d3bf13a17827381a15a4730722685a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/342520
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41898}
The DcSctpTransport will soon use field trials to conditionally enable
some options.
And overall, there is a migration project to start using the Environment
and this CL is in that direction, also setting the boundary; The dcSCTP
library should not depend on it. But the transport is allowed to.
Bug: webrtc:14997
Change-Id: I1f3c2c0d8dd7bdc698dd1d58bde7651b682bcba4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/341480
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41872}
Instead of using PacketTransportInternal::SignalReadPacket.
Bug: webrtc:15368
Change-Id: Icdc2d7f85df6db944f0ba0232891e6c5a8986a66
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/340440
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41823}
This avoids a couple of layers of error code conversion, reduces
dependency on cricket error types and allows us to preserve error
information from dcsctp. Along the way remove SendDataResult.
Bug: none
Change-Id: I1ad18a8f0b2fb181745b19c49f36f270708720c0
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/298305
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#39619}
This struct only contains two member variables now and there isn't
much value added by having it.
Low-Coverage-Reason: No change in coverage, CL modifies uncovered RTC_LOG lines.
Bug: none
Change-Id: I924d450f4c8f8e49b1cfeabaebee9fd5235a90cc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/297360
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#39563}
Instead there are direct member variables for the various relevant
states, some weren't needed, some can be const but the `id` member
in particular needs special handling and can't be const.
For dealing with the stream id, we now have SctpSid. A class that does range validation, checks thread safety, handles the special `-1` case (for what's essentially an unsigned 16 bit int). Using a special type
for this also has the effect that range checking happens more
consistently (although I'm not modifying the structs in api/).
With upcoming steps of avoiding thread hops, the ID may need to
migrate to the network thread, which the thread checks will help with.
Along the way, update SctpSidAllocator to use flat_set instead of std::set and moving some of the sctp data channel code to the cc file
to help with more accurately tracking code coverage.
Bug: webrtc:11547
Change-Id: Iea6e7647ab8f93052044c5afbcc449115206b4e9
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/296444
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#39539}
In rare cases, it is possible to queue a call to SendData from the signaling
thread on a channel being closed or already closed in the network thread.
By keeping track of currently open streams, we avoid sending messages
with a stream id of channels that the other side already considers closed
and has already reused for a new channel.
This caused rare messages to be delivered on the wrong data channel if
a message was quickly sent, channel closed and a new one reopened.
Bug: webrtc:14277
Change-Id: If35fed8d12d5d2c18cdc6601085d8b632c37a0ba
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/272624
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#37880}
The state machine for handling resets couldn't handle resets
happening from both sides at the same time.
Bug: webrtc:13994
Change-Id: I2c268e54f4c5c9858913faef91ff00f6af956e99
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/261305
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36799}
When an SCTP stream is closing, a stream reset needs
to be sent from both ends.
The remote was not sending a stream reset and quickly
opening another stream with the same StreamID could
cause SCTP errors.
Bug: webrtc:13994
Change-Id: I3abc74ddc88b3fcf7e6495d76e7d77f52280b5d1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/260922
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36773}
The variable instance_count might be accessed from multiple threads when
different PeerConnectionFactory objects are used, which may create
multiple network threads. This is a pattern mostly noticed in tests.
This fixes issues with logging when run under TSAN, it should not have
any production impact.
Bug: chromium:1243702
Change-Id: Iab1412a7907545811a309cab27a3ae23b4718606
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/251983
Auto-Submit: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Victor Boivie <boivie@google.com>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36045}
To disable dcSCTP and fallback to usrsctp, you can use the field trial
WebRTC-DataChannel-Dcsctp/Disabled/
Also remove a hidden no-break space in dcSCTP logging causing issues in
some log parsing.
Bug: chromium:1243702
Change-Id: I46136a8913a6d803a3c63c710f3ed29523e4d773
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/251867
Auto-Submit: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Victor Boivie <boivie@google.com>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36027}
To disable dcSCTP and fallback to usrsctp, you can use the field trial
WebRTC-DataChannel-Dcsctp/Disabled/
Bug: chromium:1243702
Change-Id: Ia90b796562245558a61481317bcded437400b045
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/251800
Auto-Submit: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36018}
Context: The timer precision of PostDelayedTask() is about to be lowered
to include up to 17 ms leeway. In order not to break use cases that
require high precision timers, PostDelayedHighPrecisionTask() will
continue to have the same precision that PostDelayedTask() has today.
webrtc::TaskQueueBase has an enum (kLow, kHigh) to decide which
precision to use when calling PostDelayedTaskWithPrecision().
See go/postdelayedtask-precision-in-webrtc for motivation and a table of
delayed task use cases in WebRTC that are "high" or "low" precision.
Most timers in DCSCTP are believed to only be needing low precision (see
table), but the delayed_ack_timer_ of DataTracker[1] is an example of a
use case that is likely to break if the timer precision is lowered (if
ACK is sent too late, retransmissions may occur). So this is considered
a high precision use case.
This CL makes it possible to specify the precision of dcsctp::Timer.
In a follow-up CL we will update delayed_ack_timer_ to kHigh precision.
[1] https://source.chromium.org/chromium/chromium/src/+/main:third_party/webrtc/net/dcsctp/rx/data_tracker.cc;l=340
Bug: webrtc:13604
Change-Id: I8eec5ce37044096978b5dd1985fbb00bc0d8fb7e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/249081
Reviewed-by: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35809}
This emphasizes the "hint" to potential external users that the
class has been deprecated.
Bug: webrtc:12339
Change-Id: Iab83481af69a505059297cce959f02b5ab649f2f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/237805
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35368}
Benchmarks will spam a lot of annoying lines because of this.
Bug: webrtc:13288
Change-Id: I1321abafb91baefcaa626a561bee58ca3f321291
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/235371
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35225}
When sending a large amount of data, the sender will want to keep the
send buffer full so that the socket can quickly drain it as its able to
put more bytes on the wire.
When trying to send a message and when the send buffer is full, an error
will be returned and the OnError callback will be triggered. In these
situations, this is an expected and handled error and should not be
logged as an error, as it causes confusion.
Bug: chromium:1258225
Change-Id: I3e1feab03f60ba5c41cc524d8d8f066445030d18
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/235201
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35204}
Prior to this commit, the SCTP association could terminate due to too
many retransmission attempts when there is a long duration of packet
loss. The RTCPeerConnection wouldn't terminate, and when the network
later recovers (possibly using a different ICE candidate), it would be a
RTCPeerConnection with media, but without DataChannels.
This commit will make the dcSCTP library never abort by itself when
there are too many retransmissions. It will also put a cap on the retry
duration so that it will do a retry every three seconds (or lower).
Bug: webrtc:13129
Change-Id: I08162ea20d6a60aa0eae2717966d9a2ddba8fc22
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232540
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35061}
In benchmarks, each log statement represent 2% of the CPU usage. RTC_LOG
is not very expensive, but not free either, and it's called for every
received and sent packet.
Bug: webrtc:12943
Change-Id: Id65baafb5e494091a3a7604687718fdd4f477d86
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/231223
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34929}
Using usrsctp_getladdrs() would sometimes be flagged by TSAN for a lock
order inversion. It was used to retrieve the "id" of the socket on the
transport.
The "id" is instead stored in the "ulp_info" parameter, which is
passed with each callback from usrsctp.
Bug: webrtc:12823
Change-Id: Ifb3d7780273a460e677526dd3a93f9365b29300c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/229000
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34784}
Before this change, there was no way for a client to indicate to the
dcSCTP library if a packet that was supposed to be sent, was actually
sent. It was assumed that it always was.
To handle temporary failures better, such as retrying to send packets
that failed to be sent when the send buffer was full, this information
is propagated to the library.
Note that this change only covers the API and adaptations to clients.
The actual implementation to make use of this information is done as a
follow-up change.
Bug: webrtc:12943
Change-Id: I8f9c62e17f1de1566fa6b0f13a57a3db9f4e7684
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/228563
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34767}
It is useful for more than just the transport.
Bug: webrtc:12961
Change-Id: Iad064c8fb707ca589a1c232e17436338fb06623d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/225543
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34451}
When the transport is terminated, if an error has occured, it will
be propagated to the channels.
When such errors can happen at the SCTP level (e.g. out of resources),
RTCError may contain an error code matching the definition at
https://www.iana.org/assignments/sctp-parameters/sctp-parameters.xhtml#sctp-parameters-24
If the m= line is rejected or removed from SDP, an error will again be sent
to the data channels, signaling their unexpected transition to closed.
Bug: webrtc:12904
Change-Id: Iea3d8aba0a57bbedb5d03f0fb6f7aba292e92fe8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/223541
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34386}
The factory allows us to isolate the implementation from users who only
need to depend directly on the public folder now.
Bug: webrtc:12614
Change-Id: Ied09cf772ed427eaf17a7b5705f587da57405640
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/220939
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34330}
