WebRTC binds to individual NICs and listens for incoming Stun packets. Sending stun through this specific NIC binding could make OS route the packet differently hence exposing non-VPN public IP.
The fix here is
1. to bind to any address (0:0:0:0) instead. This way, the routing will be the same as how chrome/http is.
2. also, remove the any all 0s addresses which happens when we bind to all 0s.
BUG=4276
R=juberti@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/39129004
Cr-Commit-Position: refs/heads/master@{#8418}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8418 4adac7df-926f-26a2-2b94-8c16560cd09d
VirtualSocketServer, when binding to any address (all 0s), will assign a unique IP address by incrementing the IP address, resulted in 0.0.0.1. However, this breaks the testing of 4276 where we bind to all 0s and expect the local address should remain all 0s.
BUG=4276
R=pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/35189004
Cr-Commit-Position: refs/heads/master@{#8370}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8370 4adac7df-926f-26a2-2b94-8c16560cd09d
PeerConnectionTest.java currently works, but only on a device with
network interfaces up. This is not a problem for desktop, but it is a
problem when running on Android devices since the devices in the lab
generally don't have network (due to the chaotic radio environment in
the device labs, devices are simply kept in flight mode).
The test does work if one modifies this line in the file
webrtc/base/network.cc:
bool ignored = ((cursor->ifa_flags & IFF_LOOPBACK) ||
IsIgnoredNetwork(*network));
If we remove the IFF_LOOPBACK clause, the test starts working on
an Android device in flight mode. This is nice - we're running the
call and packets interact with the OS network stack, which is good
for this end-to-end test. We can't just remove the clause though since
having loopback is undesirable for everyone except the test (right)?
so we need to make this behavior configurable.
This CL takes a stab at a complete solution where we pass a boolean
all the way through the Java PeerConnectionFactory down to the
BasicNetworkManager. This comes as a heavy price in interface
changes though. It's pretty out of proportion, but fundamentally we
need some way of telling the network manager that it is on Android
and in test mode. Passing the boolean all the way through is one way.
Another way might be to put the loopback filter behind an ifdef and
link a custom libjingle_peerconnection.so with the test. That is hacky
but doesn't pollute the interfaces. Not sure how to solve that in GYP
but it could mean some duplication between the production and
test .so files.
It would have been perfect to use flags here, but then we need to
hook up gflags parsing to some main() somewhere to make sure the
flag gets parsed, and make sure to pass that flag in our tests.
I'm not sure how that can be done.
Making the loopback filtering conditional is exactly how we solved the
equivalent problem in content_browsertests in Chrome, and it worked
great.
That's all I could think of.
BUG=4181
R=perkj@webrtc.org, pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/36769004
Cr-Commit-Position: refs/heads/master@{#8344}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8344 4adac7df-926f-26a2-2b94-8c16560cd09d
Like PlatformThreadId, this type is borrowed from Chromium.
The difference between the two is that PlatformThreadRef is pthread_t on posix platforms.
On Windows PlatformThreadRef and PlatformThreadId are the same thing.
The reason for this switch is pretty crazy. On Chromium's "Mac 10.9 dbg" bot,
we have been seeing the following code:
ThreadCheckerImpl::ThreadCheckerImpl() : valid_thread_(CurrentThreadId()) {
fprintf(stderr, "*** valid=%d\n", valid_thread_);
valid_thread_ = CurrentThreadId();
fprintf(stderr, "*** valid after=%d\n", valid_thread_);
}
print this:
*** valid=946872320
*** valid after=5647
This is for the same thread checker instance.
What's worse is that printing out what CurrentThreadId was returning, yielded that it was always returning 5647.
After switching over to pthread_t on Mac, this stopped happening.
So, to remove the current hack, reinstate the class on Mac and take a look at the next problem, I'm switching to pthread_t.
Really looking forward to truly getting to the bottom of this.
Tbr-ing since the build is essentially broken (we can't roll).
TBR=pbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/37199004
Cr-Commit-Position: refs/heads/master@{#8283}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8283 4adac7df-926f-26a2-2b94-8c16560cd09d
I'm reverting the patch due to compilation issues. It would be great if we could make sure Chromium is ready for the patch before we land it in WebRTC.
As is, we're trying to roll webrtc into Chromium and we can't (this is not the only reason though). I might reland this after the roll, depending on how that goes though.
Here's an example failure:
e:\b\build\slave\win_gn\build\src\jingle\glue\channel_socket_adapter_unittest.cc(77) : error C2259: 'jingle_glue::MockTransportChannel' : cannot instantiate abstract class
due to following members:
'bool cricket::TransportChannel::GetSslCipher(std::string *)' : is abstract
e:\b\build\slave\win_gn\build\src\third_party\webrtc\p2p\base\transportchannel.h(107) : see declaration of 'cricket::TransportChannel::GetSslCipher'
ninja: build stopped: subcommand failed.
> This CL adds an API to the SSL stream adapters and transport channels to get the SSL cipher that was negotiated with the remote peer.
>
> BUG=3976
> R=davidben@chromium.org, juberti@webrtc.org, pthatcher@webrtc.org
>
> Review URL: https://webrtc-codereview.appspot.com/26009004TBR=pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40689004
Cr-Commit-Position: refs/heads/master@{#8282}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8282 4adac7df-926f-26a2-2b94-8c16560cd09d
Failed on Linux_Memcheck bot.
http://chromegw/i/client.webrtc/builders/Linux%20Memcheck/builds/3182
> VirtualSocketServer out-of-order issue with closing TCP sockets
>
> https://webrtc-codereview.appspot.com/41449004 added a TURN TCP
> allocation release test which was disabled as it triggered an assert
> in the turnserver.
>
> This was caused by VirtualSockerServer delivering the last TCP packet
> after closing the connection. Calling
> VirtualSocketServer::SendTcp
> and
> VirtualSocket::Close
> from TestTurnTCPReleaseAllocation led to the following order of
> messages in VirtualSocket::OnMessage:
> MSG_ID_DISCONNECT
> MSG_ID_PACKET
>
> This is out of order and triggers an assert in turnserver.cc since the
> socket from which the message arrives has already been discarded,
> subsequently breaking the test.
>
> In VirtualSocketServer::Disconnect the MSG_ID_DISCONNECT is posted to the
> msg_queue immediately, thus getting ahead of any (slightly delayed)
> actual packets.
>
> Maybe PostAt(network_delay_ + 1, ...) would be better?
>
> Re-enables TestTurnTCPReleaseAllocation.
>
> BUG=
> R=juberti@webrtc.org
>
> Review URL: https://webrtc-codereview.appspot.com/34759004TBR=pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/38979004
Cr-Commit-Position: refs/heads/master@{#8280}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8280 4adac7df-926f-26a2-2b94-8c16560cd09d
https://webrtc-codereview.appspot.com/41449004 added a TURN TCP
allocation release test which was disabled as it triggered an assert
in the turnserver.
This was caused by VirtualSockerServer delivering the last TCP packet
after closing the connection. Calling
VirtualSocketServer::SendTcp
and
VirtualSocket::Close
from TestTurnTCPReleaseAllocation led to the following order of
messages in VirtualSocket::OnMessage:
MSG_ID_DISCONNECT
MSG_ID_PACKET
This is out of order and triggers an assert in turnserver.cc since the
socket from which the message arrives has already been discarded,
subsequently breaking the test.
In VirtualSocketServer::Disconnect the MSG_ID_DISCONNECT is posted to the
msg_queue immediately, thus getting ahead of any (slightly delayed)
actual packets.
Maybe PostAt(network_delay_ + 1, ...) would be better?
Re-enables TestTurnTCPReleaseAllocation.
BUG=
R=juberti@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/34759004
Cr-Commit-Position: refs/heads/master@{#8271}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8271 4adac7df-926f-26a2-2b94-8c16560cd09d
In order to figure out the issue with the Mac 10.9 debug bot, this patch disables the ThreadChecker class on Mac in debug builds. For diagnostic purposes, it instead prints out when there's a thread mismatch. I'm also adding a DCHECK in case fetching the current thread id ever returns 0.
R=magjed@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40679004
Cr-Commit-Position: refs/heads/master@{#8269}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8269 4adac7df-926f-26a2-2b94-8c16560cd09d
This is the same change as already made for Windows:
https://webrtc-codereview.appspot.com/37069004/
* Remove "dead" and "alive" variables.
* Remove critical section
* Remove implementation of SetNotAlive()
* Always set thread name
* Add thread checks for correct usage.
* Changed AudioDeviceMac to create/start/stop/delete thread objects for playout and recording, inside the respective start and stop method. The reason for this is because the AudioDeviceMac instance is currently being created on one thread and the above Start/Stop methods are being called on a different thread. So, my change makes creation, start/stop, deletion of the thread objects always happen on the same thread.
I'm making CurrentThreadId() in rtc_base_approved more visible so that it can be used from there instead of inside webrtc. Down the line we will have more thread concepts in rtc_base_approved, so I put a TODO for myself to move this functionality to there once we do.
R=henrika@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40599004
Cr-Commit-Position: refs/heads/master@{#8235}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8235 4adac7df-926f-26a2-2b94-8c16560cd09d
The latter file was more up-to-date. The files are now identical
with the following exceptions:
* The namespace used (rtc vs. webrtc).
* The name of the include guard.
* base/scoped_ptr.h still has two extra methods, accept() and use().
* base/scoped_ptr.h still includes webrtc/base/common.h even though
it doesn't need it itself, since several .cc files expect to get
it for free by incuding base/scoped_ptr.h. This is of course bad
manners, and the "unused" include will be removed in a future CL.
A later CL will remove system_wrappers/interface/scoped_ptr.h.
R=andrew@webrtc.org
Committed: https://code.google.com/p/webrtc/source/detail?r=8147
And reverted again, because out-of-tree code using this file was defining nullptr to 0: https://code.google.com/p/webrtc/source/detail?r=8149
Review URL: https://webrtc-codereview.appspot.com/36919004
Cr-Commit-Position: refs/heads/master@{#8196}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8196 4adac7df-926f-26a2-2b94-8c16560cd09d
Currently, in ipc_network_manager.cc, the UMA WebRTC.PeerConnection.IPv4Interfaces and its IPv6
counter part counts the addresses, instead of the interfaces as when
chromium delivers available networks to WebRTC, each address is wrapped
inside an individual network object.
The plan is to replace the current MergeNetworkList with the new one once it's rolled into chromium.
BUG=
R=pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/36779004
Cr-Commit-Position: refs/heads/master@{#8188}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8188 4adac7df-926f-26a2-2b94-8c16560cd09d
To do this, I'm removing ThreadChecker's dependency on the 'Thread' class, so that the checker works with any thread and doesn't rely on TLS.
Also simplifying CriticalSection's implementation on Windows since a critical section on Windows already knows what thread currently owns the lock.
BUG=
R=mflodman@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40539004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8151 4adac7df-926f-26a2-2b94-8c16560cd09d
The latter file was more up-to-date. The files are now identical
with the following exceptions:
* The namespace used (rtc vs. webrtc).
* The name of the include guard.
* base/scoped_ptr.h still has two extra methods, accept() and use().
* base/scoped_ptr.h still includes webrtc/base/common.h even though
it doesn't need it itself, since several .cc files expect to get
it for free by incuding base/scoped_ptr.h. This is of course bad
manners, and the "unused" include will be removed in a future CL.
A later CL will remove system_wrappers/interface/scoped_ptr.h.
R=andrew@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/36919004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8147 4adac7df-926f-26a2-2b94-8c16560cd09d
This patch basically deletes webrtc/base/template_util.h (which is the
more outdated copy, although there are only cosmetical differences)
and moves webrtc/system_wrappers/source/template_util.h to take its
place.
The reunified header uses the rtc namespace like the old
webrtc/base/template_util.h, rather than the webrtc namespace like
webrtc/system_wrappers/source/template_util.h.
R=aluebs@webrtc.org, andrew@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/38579004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8050 4adac7df-926f-26a2-2b94-8c16560cd09d
