It was possible that "A < B" and "B < A" both evaluated to true.
This manifested as an assert on Windows, and a memory leak on Linux.
Note that the concept of "less than" is meaningless for this object.
The operator is only needed so the object can be used as a key in an
std::map.
BUG=webrtc:6068
R=honghaiz@webrtc.org, kjellander@webrtc.org, skvlad@webrtc.org
Review URL: https://codereview.webrtc.org/2187913002 .
Cr-Commit-Position: refs/heads/master@{#13598}
If a port is not used by any channel and if it has no connection for 30
seconds, it will be removed.
Note, as long as a port is used by a transport channel, it will be kept
even if it does not have any connection. This will be beneficial to
continual gathering because new connections can be created in the future
when network changes.
BUG=
R=pthatcher@webrtc.org, zhihuang@webrtc.org
Review URL: https://codereview.webrtc.org/2171183002 .
Cr-Commit-Position: refs/heads/master@{#13567}
This change makes WebRTC no longer stop sending video when we receive an
EWOULDBLOCK error from the operating system. This was previously
causing calls on a slow link (where the first hop is slow) to rapidly
oscillate between starting and stopping video.
We still do need to stop sending packets if there is no known good
connection we can use for that. We used to generate a synthetic
EWOULDBLOCK error in that case. This CL replaces it with a different
code (ENOTCONN); EWOULDBLOCK no longer stops the stream but ENOTCONN
does.
I've updated all the places where we seemed to be generating EWOULDBLOCK
for reasons other than some buffer been full; please give it a thorough
look in case I missed something.
R=pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2192963002 .
Cr-Commit-Position: refs/heads/master@{#13566}
If all connections on a port is destroyed, it will schedule an event
to check if it is dead after a timeout. Previously if a new connection
is created but destroyed before the event is fired, it will destroy the
port. With this change, we will not destoy it until it times out again
after the last created connection is destroyed.
BUG=
R=pthatcher@webrtc.org, zhihuang@webrtc.org
Review URL: https://codereview.webrtc.org/2184013003 .
Cr-Commit-Position: refs/heads/master@{#13563}
We thought we could safely remove this, but older versions of Chrome
don't do role conflict resolution properly, so it's actually not safe
to yet.
BUG=628676
Review-Url: https://codereview.webrtc.org/2152963003
Cr-Commit-Position: refs/heads/master@{#13492}
If the currently selected connection becomes not receiving and if a backup connection
becomes strong first, we will not switch the connection until X milliseconds is passed
but the selected connection is still not receiving and the backup connection is still receiving. This will prevent the connection switching from happening too frequently.
BUG=
Review-Url: https://codereview.webrtc.org/2143653005
Cr-Commit-Position: refs/heads/master@{#13480}
The DtlsTransportChannel wasn't properly handling the
P2PTransportChannel becoming writable before receiving a remote
fingerprint; it wasn't starting the DTLS handshake whenever this
happened.
Review-Url: https://codereview.webrtc.org/2140283002
Cr-Commit-Position: refs/heads/master@{#13469}
If the test TURN server received two allocate requests from the same
address, it was replacing the old allocation but not deleting it.
Also switching to std::unique_ptr to make it less likely for this to
pop up again.
Review-Url: https://codereview.webrtc.org/2114063002
Cr-Commit-Position: refs/heads/master@{#13449}
we will periodically check if any network does not have any connection on it and if yes, attempt to re-gather on those networks.
BUG=
R=pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2025573002 .
Cr-Commit-Position: refs/heads/master@{#13367}
When the flag prune_turn_ports is set, When a high-priority turn port becomes available, it will prune low-priority ones. The pruned port will not be used for creating connections locally and its candidates will not be sent over to the remove side (unless they have been sent before being pruned).
This effectively reduces the number of TURN candidates and connections created by TURN ports.
BUG=
R=deadbeef@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2093623004 .
Committed: https://crrev.com/17aac053f585e892114974d2eb248e05ad37f973
Cr-Original-Commit-Position: refs/heads/master@{#13335}
Cr-Commit-Position: refs/heads/master@{#13354}
Reason for revert:
Breaks Win32/Win64 Debug bots in client.webrtc waterfall
Original issue's description:
> Add config to prune low-priority TURN ports for creating connections
> When the flag prune_turn_ports is set, When a high-priority turn port becomes available, it will prune low-priority ones. The pruned port will not be used for creating connections locally and its candidates will not be sent over to the remove side (unless they have been sent before being pruned).
>
> This effectively reduces the number of TURN candidates and connections created by TURN ports.
>
> BUG=
> R=deadbeef@webrtc.org, pthatcher@webrtc.org
>
> Committed: https://crrev.com/17aac053f585e892114974d2eb248e05ad37f973
> Cr-Commit-Position: refs/heads/master@{#13335}
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org,honghaiz@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review-Url: https://codereview.webrtc.org/2111663003
Cr-Commit-Position: refs/heads/master@{#13342}
When the flag prune_turn_ports is set, When a high-priority turn port becomes available, it will prune low-priority ones. The pruned port will not be used for creating connections locally and its candidates will not be sent over to the remove side (unless they have been sent before being pruned).
This effectively reduces the number of TURN candidates and connections created by TURN ports.
BUG=
R=deadbeef@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2093623004 .
Cr-Commit-Position: refs/heads/master@{#13335}
Previously, we were starting a periodic timer when the local
description was set. The first connection may be created at any
time after this happens, so after creating the first connection, we
need to wait until that timer next fires before sending a ping.
Now we just start that timer (and send the first ping) immediately
after the first connection becomes pingable.
This CL also removes the "Connect" method. The only vestigal
effect of this method was to start the periodic timer, which is
now not needed since it happens automatically.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2099563004 .
Cr-Commit-Position: refs/heads/master@{#13331}
The "should I simulate EWOULDBLOCK?" determination now happens
solely in P2PTransportChannel. This also fixes a bug where the
"last packet id" was set even if no packet was sent.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2099783002 .
Cr-Commit-Position: refs/heads/master@{#13307}
The fake clock has a few advantages:
1. It lets use verify that operations take the expected number of
round trips.
2. It makes the tests faster by letting us remove the equivalent
of "Sleep(500)" all over the tests.
3. It makes the tests less flaky, because sometimes sleeping for
500ms or waiting for 1s is not enough.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2097793003 .
Cr-Commit-Position: refs/heads/master@{#13304}
Reason for revert:
Didn't intend to land yet. Chromium CL still needed.
Original issue's description:
> Add virtual Initialize methods to PortAllocator and NetworkManager.
>
> This will allow PeerConnection to handle hopping to the right thread
> and doing thread-specific initialization for the PortAllocator.
> This eliminates a required thread-hop for whatever is passing the
> PortAllocator into CreatePeerConnection.
>
> BUG=617648
> R=pthatcher@webrtc.org, skvlad@webrtc.org
>
> Committed: https://crrev.com/a6bdb0990a659ff9e7c4374f5033a6bcc4fbfb21
> Cr-Commit-Position: refs/heads/master@{#13283}
TBR=pthatcher@webrtc.org,skvlad@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=617648
Review-Url: https://codereview.webrtc.org/2092023004
Cr-Commit-Position: refs/heads/master@{#13284}
This will allow PeerConnection to handle hopping to the right thread
and doing thread-specific initialization for the PortAllocator.
This eliminates a required thread-hop for whatever is passing the
PortAllocator into CreatePeerConnection.
BUG=617648
R=pthatcher@webrtc.org, skvlad@webrtc.org
Review URL: https://codereview.webrtc.org/2097653002 .
Cr-Commit-Position: refs/heads/master@{#13283}
Reason for revert:
It turns out this revert was not necessary because the connection-state mapping for turn-turn connections was not done in connection.
Original issue's description:
> Revert of Revert "Revert of Update the BWE when the network route changes. (patchset #5 id:180001 of https://… (patchset #5 id:120001 of https://codereview.webrtc.org/2041593002/ )
>
> Reason for revert:
> ReadyToSendMedia did not consider the new presumed_writable state.
>
> Original issue's description:
> > Revert "Revert of Update the BWE when the network route changes. (patchset #5 id:180001 of https://codereview.webrtc.org/2000063003/ )"
> >
> > This reverts commit 72d41aa6da94dacb8a8464d1abd4ca7d1afffc65.
> >
> > New change made:
> > Do not reset the BWE when the new network route is not ready to send media.
> >
> > BUG=
> > R=pthatcher@webrtc.org, stefan@webrtc.org
> >
TBR=pthatcher@webrtc.org,stefan@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review-Url: https://codereview.webrtc.org/2094863003
Cr-Commit-Position: refs/heads/master@{#13282}
We now use a single rule to determine connection switch on the controlled side. The rule is to select the new best connection based on the following order:
1. writable/receiving/connected state.
2. nominated
3. last time receiving data packet.
4. priority.
5. latency (rtt)
BUG=
R=deadbeef@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2069493002 .
Cr-Commit-Position: refs/heads/master@{#13274}
When creating connections on turn port, check whether the local and remote candidates have the same IP address family, instead of checking the address family of the local socket against the remote candidate.
BUG=5871
R=deadbeef@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2083803002 .
Cr-Commit-Position: refs/heads/master@{#13269}
This will allow media to be sent on these pairs before a binding
response is received, shortening call setup time. However, this is only
possible if the TURN servers don't require CreatePermission when
communicating with each other.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
NOTRY=True
Review-Url: https://codereview.webrtc.org/2063823008
Cr-Commit-Position: refs/heads/master@{#13268}
Reason for revert:
The Webrtc waterfall indicates that this revert is not necessary.
Original issue's description:
> Revert of Do not delete a connection in the turn port with permission error or refresh error. (patchset #6 id:260001 of https://codereview.webrtc.org/2068263003/ )
>
> Reason for revert:
> It broke webrtc builds.
>
> Original issue's description:
> > Do not delete a connection in the turn port with permission error, refresh error, or binding error.
> >
> > Even if those error happened, the connection may still be able to receive packets for a while.
> > If we delete the connections, all packets arriving will be dropped.
> >
> > BUG=webrtc:6007
> > R=deadbeef@webrtc.org, pthatcher@webrtc.org
> >
> > Committed: https://crrev.com/3d77deb29c15bfb8f794ef3413837a0ec0f0c131
> > Cr-Commit-Position: refs/heads/master@{#13262}
>
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
> # Not skipping CQ checks because original CL landed more than 1 days ago.
> BUG=webrtc:6007
>
> Committed: https://crrev.com/3159ffae6b1d5cba2ad972bd3d8074ac85f2c7f9
> Cr-Commit-Position: refs/heads/master@{#13265}
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:6007
Review-Url: https://codereview.webrtc.org/2090073003
Cr-Commit-Position: refs/heads/master@{#13266}
Reason for revert:
It broke webrtc builds.
Original issue's description:
> Do not delete a connection in the turn port with permission error, refresh error, or binding error.
>
> Even if those error happened, the connection may still be able to receive packets for a while.
> If we delete the connections, all packets arriving will be dropped.
>
> BUG=webrtc:6007
> R=deadbeef@webrtc.org, pthatcher@webrtc.org
>
> Committed: https://crrev.com/3d77deb29c15bfb8f794ef3413837a0ec0f0c131
> Cr-Commit-Position: refs/heads/master@{#13262}
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:6007
Review-Url: https://codereview.webrtc.org/2090833002
Cr-Commit-Position: refs/heads/master@{#13265}
Reason for revert:
Breaking webrtc builder.
Original issue's description:
> Adding IceConfig option to assume TURN/TURN candidate pairs will work.
>
> This will allow media to be sent on these pairs before a binding
> response is received, shortening call setup time. However, this is only
> possible if the TURN servers don't require CreatePermission when
> communicating with each other.
>
> R=honghaiz@webrtc.org, pthatcher@webrtc.org
>
> Committed: https://crrev.com/8e6134eae4117a239de67c9a9dae8f5e3235d803
> Cr-Commit-Position: refs/heads/master@{#13263}
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
Review-Url: https://codereview.webrtc.org/2090823002
Cr-Commit-Position: refs/heads/master@{#13264}
This will allow media to be sent on these pairs before a binding
response is received, shortening call setup time. However, this is only
possible if the TURN servers don't require CreatePermission when
communicating with each other.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2063823008 .
Cr-Commit-Position: refs/heads/master@{#13263}
Even if those error happened, the connection may still be able to receive packets for a while.
If we delete the connections, all packets arriving will be dropped.
BUG=webrtc:6007
R=deadbeef@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2068263003 .
Cr-Commit-Position: refs/heads/master@{#13262}
If an actual peer reflexive candidate was created (and not one that
would just be replaced by a different candidate later), we weren't
setting the generation value. This means that new-generation prflx
candidate pairs weren't being prioritized above the cross-generation
pairs, or above relay<->relay pairs.
R=honghaiz@webrtc.org, pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/2086793002 .
Cr-Commit-Position: refs/heads/master@{#13259}
Previously, if the ICE role changed, SetIceRole was only called on
the ports from the most recent ICE generation. However, STUN pings
may still be sent and received by older generation ports, so they
should receive an updated role as well.
This was previously triggering an ASSERT, because a P2PTransportChannel
expects the ICE role of each of its ports to match its own role.
Committed: https://crrev.com/370544594e18deb7f560f961295c8cf3f0a679f1
Review-Url: https://codereview.webrtc.org/2053043003
Cr-Original-Commit-Position: refs/heads/master@{#13226}
Cr-Commit-Position: refs/heads/master@{#13247}
Previously they were only being updated for connections using the
most current "generation" of ports. This results in the older-
generation prflx candidate pair being prioritized above newer-
generation candidate pairs.
Review-Url: https://codereview.webrtc.org/2087713002
Cr-Commit-Position: refs/heads/master@{#13245}
Writable connections are pinged at a slower rate.
The function IsPingable will filter out the writable connections.
The interval for slower ping rate by default is WRITABLE_CONNECTION_PING_INTERVAL(2500ms) and can be set with the configuration.
BUG=webrtc:1161
Committed: https://crrev.com/8f7a5aad55a64f0d81b6436a22ffbdfcdcde91e0
Review-Url: https://codereview.webrtc.org/1944003002
Cr-Original-Commit-Position: refs/heads/master@{#12736}
Cr-Commit-Position: refs/heads/master@{#13241}
Reason for revert:
Speculative revert: breaks video quality tests on Win and Mac (???): https://build.chromium.org/p/chromium.webrtc.fyi/builders/Mac%20Tester/builds/31209
Original issue's description:
> Update ICE role on all ports, not just ones used for new connections.
>
> Previously, if the ICE role changed, SetIceRole was only called on
> the ports from the most recent ICE generation. However, STUN pings
> may still be sent and received by older generation ports, so they
> should receive an updated role as well.
>
> This was previously triggering an ASSERT, because a P2PTransportChannel
> expects the ICE role of each of its ports to match its own role.
>
> Committed: https://crrev.com/370544594e18deb7f560f961295c8cf3f0a679f1
> Cr-Commit-Position: refs/heads/master@{#13226}
TBR=pthatcher@webrtc.org,honghaiz@webrtc.org,deadbeef@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review-Url: https://codereview.webrtc.org/2078423004
Cr-Commit-Position: refs/heads/master@{#13240}
Previously, if the ICE role changed, SetIceRole was only called on
the ports from the most recent ICE generation. However, STUN pings
may still be sent and received by older generation ports, so they
should receive an updated role as well.
This was previously triggering an ASSERT, because a P2PTransportChannel
expects the ICE role of each of its ports to match its own role.
Review-Url: https://codereview.webrtc.org/2053043003
Cr-Commit-Position: refs/heads/master@{#13226}
Every message will now be traced with the location from which it was
posted, including function name, file and line number.
This CL also writes a normal LOG message when the dispatch took more
than a certain amount of time (currently 50ms).
This logging should help us identify messages that are taking
longer than expected to be dispatched.
R=pthatcher@webrtc.org, tommi@webrtc.org
Review URL: https://codereview.webrtc.org/2019423006 .
Cr-Commit-Position: refs/heads/master@{#13104}
