When a socket is shutting down, either explicitly by the ULP calling
Shutdown(), or when the socket receives a SHUTDOWN chunk, the socket
should send all outstanding data and when all is sent and acked,
_then_ it should continue the shutdown protocol.
As it currently doesn't calculate correctly when all data has been sent,
as NACKED chunks are not included in what it believes is remaining in
the retransmission queue, it will shut down prematurely and may go back
to a previous state (ShutdownPending) from ShutdownSent or
ShutdownAckSent.
This is a workaround that just avoids the illegal state transition as
that puts the socket in an inconsistent state. The bug is merely
theoretical as WebRTC doesn't currently gracefully shut down a SCTP
socket, but just terminates the DTLS transport.
As TODOs mention, this will be fixed correctly a bit later.
Bug: webrtc:12739
Change-Id: Ibde2acc3a6aca701ac178d6181028404d470a5d5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/218340
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33982}
Historically the PacketBuffer used a callback for assembled frames, and because of that RtpPacketInfos were piped through it even though they didn't have anything to do with the PacketBuffer.
With this CL RtpPacketInfos are stored in the RtpVideoStreamReceiver(2) instead.
Bug: webrtc:12579
Change-Id: Ia6285b59e135910eee7234b89b23425bb0fc0d2b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215320
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33980}
Before this CL, the RemoteEstimatorProxy used a std::map to track which
arrival time a packet with a certain sequence number was received at.
While this works, it's fairly slow as most manipulations and insertions
are O(log(N)) and there were quite many of them.
By taking advantage that sequence numbers generally are received in
sequence, recording a packet is now amortized O(1). Also other
operations such as creating the periodic feedback reports, are also
much faster as it previously was done by searching quite a few times
in that map.
In highly loaded Media Servers, RemoteEstimatorProxy's usage of
std::map attributes to around 0.52% CPU.
Bug: webrtc:12689
Change-Id: I3dd58105f9fbfb111f176833cd4aa6b040c0e01d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217388
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33979}
While it's not strictly defined, the expectation is that sending a
message with a lifetime parameter set to zero (0) ms should allow it to
be sent if it can be sent without being buffered. If it can't be
directly sent, it should be discarded.
This is initial support for it. Small messages can now be delivered fine
if they are not to be buffered, but fragmented messages could be partly
sent (if this fills up the congestion window), which means that the
message will then fail to be sent whenever the congestion window frees
up again. It would be better to - at a higher level - realize early that
the message can't be sent in full, and discard it without sending
anything. But that's an optimization that can be done later.
A few off-by-one errors were found when strictly defining that the
message is alive during its entire lifetime. It will expire just _after_
its lifetime.
Sending messages with a lifetime of zero may not supported in all
libraries, so a workaround would be to set a very small timeout instead,
which is tested as well.
Bug: webrtc:12614
Change-Id: I9a00bedb639ad7b3b565b750ef2a49c9020745f1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217562
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33977}
Keeping just the header doesn't save memory because header is taken as slice
of the original packet (and thus keeps a reference to the buffer containing
full packet)
Keeping full packet is simpler and avoid extra unused buffer created during
RtpPacket default contruction
Bug: b/187593466
Change-Id: I78d7201d110092fc039203e1caa2fb9c3afbc079
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/218161
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33974}
Due to a limit socket send buffer, it's quite easy to fill it up when
using exponential slow start, which results in dropping a lot of packets
and having to retransmit those.
Disabling this, to align it to how SCTP normally behaves, and then try
to stabilize it later. With SCTP slow start, it will increase with one
MTU for each RTT when there is no packet loss. Even this mode will
experience packet loss, but not as much will be lost, and it will
stabilize quicker.
Bug: webrtc:12614
Change-Id: Ibc484b19b7e708fe5bd837bbef178a2f69b7211f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/218203
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33969}
This CL enables the RTC_ENABLE_WIN_WGC build flag, which introduces a
dependency on the Win10 SDK v10.0.17763. There is no change in behavior
from enabling this flag.
Consumers of WebRTC that use an older version of the Win10 SDK will
see errors similar to:
fatal error: 'windows.graphics.capture.interop.h' file not found
fatal error: 'windows.graphics.capture.h' file not found
They should upgrade to this or a newer version (Chromium requires,
and thus WebRTC recommends, v10.0.19041). You can find instructions
here:
https://chromium.googlesource.com/chromium/src/+/master/docs/windows_build_instructions.md
Alternatively, consumers can disable this build flag in their
downstream copies of the WebRTC repo.
Bug: webrtc:9273
Change-Id: Ic6bf3ef3e69b8ba0e4022e07832fa66b6bcc6740
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215244
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Austin Orion <auorion@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#33968}
cricket::SendDataParams is replaced by webrtc::SendDataParams.
cricket::DataMessageType is replaced by webrtc::DataMessageType.
The sid member from cricket::SendDataParams is now passed as an argument
to functions that used one when necessary.
Bug: webrtc:7484
Change-Id: Ia4a89c9651fb54ab9a084a6098d49130b6319e1b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217761
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33966}
Applications should use CreatePeerConnectionOrError instead.
Moved fallback implementations of CreatePeerConnection into the
api/peer_connection_interface.h file, so that we do not have to
declare these methods in the proxy.
Bug: webrtc:12238
Change-Id: I70c56336641c2a108b68446ae41f43409277a584
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217762
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33964}
When the shutdown timer has expired, the socket will abort/close and the
TCB is not valid after InternalClose.
Bug: webrtc:12614
Change-Id: I09a94a049f0cda4577225dd9c80a92a8ec7e0423
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217767
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33956}
Pending messages on network thread for MediaChannel, will be dropped
when the MediaChannel object is deleted (without blocking).
Remove MessageHandler inheritance from Channel since Post-ing to the
network thread has been removed from there.
Copy/pasted code for SendRtp/SendRtcp in WebRtcVideoChannel and
WebRtcVoiceMediaChannel consolidated in MediaChannel.
Bug: webrtc:11993
Change-Id: I05320eb7f86b98adba50ca5eb8b76b78f4111263
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217720
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33955}
It doesn't make sense to use negative values or 0 to disable the
feature, so we use an optional int value.
Values bigger than 65535 are clamped down.
Bug: webrtc:12730
Change-Id: I6bd9cd92f7d0a70a78cf5a7c91dca52c28d08ba1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217760
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33954}
This CL mostly adds plumbing to get awareness of the network thread
to the media channel classes. Currently this pointer is only used
to DCHECK that `SetInterface` for the `NetworkInterface` pointer, is
called on the network thread. Follow up changes will establish that
most of the methods are called on the network thread and the mutex
in the MediaChannel base class, can be removed.
Most of the changes in the CL are in channel_unittest.cc. They're mostly
around updating the tests to incorporate the network thread in ways
that reflect how the classes are used in production. Another change is
to use accessor methods for the media channel instances instead of
caching potentially dangling pointers.
Bug: webrtc:11993
Change-Id: I8e2ed1bc23724e238554dbce386789d69660f7e4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217682
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33951}
This reverts commit 793bac569fdf1be16cbf24d7871d20d00bbec81b.
Reason for revert: rare compilation error fixed
Original change's description:
> Revert "Refactor the PlatformThread API."
>
> This reverts commit c89fdd716c4c8af608017c76f75bf27e4c3d602e.
>
> Reason for revert: Causes rare compilation error on win-libfuzzer-asan trybot.
> See https://ci.chromium.org/p/chromium/builders/try/win-libfuzzer-asan-rel/713745?
>
> Original change's description:
> > Refactor the PlatformThread API.
> >
> > PlatformThread's API is using old style function pointers, causes
> > casting, is unintuitive and forces artificial call sequences, and
> > is additionally possible to misuse in release mode.
> >
> > Fix this by an API face lift:
> > 1. The class is turned into a handle, which can be empty.
> > 2. The only way of getting a non-empty PlatformThread is by calling
> > SpawnJoinable or SpawnDetached, clearly conveying the semantics to the
> > code reader.
> > 3. Handles can be Finalized, which works differently for joinable and
> > detached threads:
> > a) Handles for detached threads are simply closed where applicable.
> > b) Joinable threads are joined before handles are closed.
> > 4. The destructor finalizes handles. No explicit call is needed.
> >
> > Fixed: webrtc:12727
> > Change-Id: Id00a0464edf4fc9e552b6a1fbb5d2e1280e88811
> > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215075
> > Commit-Queue: Markus Handell <handellm@webrtc.org>
> > Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> > Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
> > Reviewed-by: Tommi <tommi@webrtc.org>
> > Cr-Commit-Position: refs/heads/master@{#33923}
>
> # Not skipping CQ checks because original CL landed > 1 day ago.
>
> TBR=handellm@webrtc.org
>
> Bug: webrtc:12727
> Change-Id: Ic0146be8866f6dd3ad9c364fb8646650b8e07419
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217583
> Reviewed-by: Guido Urdaneta <guidou@webrtc.org>
> Reviewed-by: Markus Handell <handellm@webrtc.org>
> Commit-Queue: Guido Urdaneta <guidou@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33936}
# Not skipping CQ checks because this is a reland.
Bug: webrtc:12727
Change-Id: Ifd6f44eac72fed84474277a1be03eb84d2f4376e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217881
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Markus Handell <handellm@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33950}
If Shutdown is called when the socket is being established and while the
connection timers are running, it will put the socket in an inconsistent
state, which is verified in debug builds.
Bug: webrtc:12614
Change-Id: I66f07d1170ac8f0ad9fd485d77d6aef4c365f150
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217765
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33949}
This makes it easier to understand which socket that experience an error
or abort. Aborts are now also logged, which was missed previously.
Bug: webrtc:12614
Change-Id: Ie5e4357b3e5450106cc6cc28c1e9578ad53d073a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217764
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33947}
As it will be heavily modified in a follow-up change, it's refactored
as a separate commit to make commits small:
* Extracted code to separate AddPacket and CullOldPackets methods
* BuildFeedbackPacket now returns a packet. In the next iteration, it
might not, so it needs to be able to decide when to increment the
packet sequence number.
* Documented some existing fields.
The follow-up is in change I3dd58105f9fbfb111f176833cd4aa6b040c0e01d.
Bug: webrtc:12689
Change-Id: I5ad0aee5a7da008f9e209f7c13bf299c12f9d1f3
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217581
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33942}
But keep option to disable it.
The fix has been tested using A/B experiment with Duo
with positive results.
Bug: webrtc:10707
Change-Id: Iaac72430bc52026ab5db9dd042c5fdc3832fdce5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217380
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Commit-Queue: Jonas Oreland <jonaso@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33941}
This reverts commit c89fdd716c4c8af608017c76f75bf27e4c3d602e.
Reason for revert: Causes rare compilation error on win-libfuzzer-asan trybot.
See https://ci.chromium.org/p/chromium/builders/try/win-libfuzzer-asan-rel/713745?
Original change's description:
> Refactor the PlatformThread API.
>
> PlatformThread's API is using old style function pointers, causes
> casting, is unintuitive and forces artificial call sequences, and
> is additionally possible to misuse in release mode.
>
> Fix this by an API face lift:
> 1. The class is turned into a handle, which can be empty.
> 2. The only way of getting a non-empty PlatformThread is by calling
> SpawnJoinable or SpawnDetached, clearly conveying the semantics to the
> code reader.
> 3. Handles can be Finalized, which works differently for joinable and
> detached threads:
> a) Handles for detached threads are simply closed where applicable.
> b) Joinable threads are joined before handles are closed.
> 4. The destructor finalizes handles. No explicit call is needed.
>
> Fixed: webrtc:12727
> Change-Id: Id00a0464edf4fc9e552b6a1fbb5d2e1280e88811
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215075
> Commit-Queue: Markus Handell <handellm@webrtc.org>
> Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
> Reviewed-by: Tommi <tommi@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#33923}
# Not skipping CQ checks because original CL landed > 1 day ago.
TBR=handellm@webrtc.org
Bug: webrtc:12727
Change-Id: Ic0146be8866f6dd3ad9c364fb8646650b8e07419
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217583
Reviewed-by: Guido Urdaneta <guidou@webrtc.org>
Reviewed-by: Markus Handell <handellm@webrtc.org>
Commit-Queue: Guido Urdaneta <guidou@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33936}
