Fixing issue with DTLS channel when using "presume writable" flag.

webrtc/p2p/base/dtlstransportchannel.cc

Index: webrtc/p2p/base/dtlstransportchannel.cc
diff --git a/webrtc/p2p/base/dtlstransportchannel.cc b/webrtc/p2p/base/dtlstransportchannel.cc
index d25de260f380693a3c8bd25312e42c1fbc3b53b7..2880200fefee9451929bbe9da237487451becb88 100644
--- a/webrtc/p2p/base/dtlstransportchannel.cc
+++ b/webrtc/p2p/base/dtlstransportchannel.cc
@@ -247,6 +247,10 @@ bool DtlsTransportChannelWrapper::SetRemoteFingerprint(
 
   if (reconnect) {
     Reconnect();
+  } else {
+    // If the underlying channel is already writable by the time we receive
+    // a remote fingerprint, we may be able to start DTLS right away.
+    MaybeStartDtls();

[skvlad, 2016/07/12 21:53:52]

The other two calls wrap this in VERIFY() - do you think this should also be
done here?

[pthatcher1, 2016/07/13 00:04:19]

If the other call is inlined, then there would just be one call to
MaybeStartDtls here, outside of the if statement.

[Taylor Brandstetter, 2016/07/13 16:17:41]

If we always want to DCHECK that this succeeds, I'll just move the DCHECK into
the function itself and have it return void.

   }
 
   return true;
@@ -675,9 +679,7 @@ void DtlsTransportChannelWrapper::OnChannelStateChanged(
 void DtlsTransportChannelWrapper::Reconnect() {
   set_dtls_state(DTLS_TRANSPORT_NEW);

[pthatcher1, 2016/07/13 00:04:19]

Maybe have a DCHECK(dtls_active_);

[Taylor Brandstetter, 2016/07/13 16:17:41]

No longer necessary. This is inlined in SetRemoteFingerprint which explicitly
checks dtls_active_.

   set_writable(false);
-  if (channel_->writable()) {
-    OnWritableState(channel_);
-  }
+  VERIFY(MaybeStartDtls());

[pthatcher1, 2016/07/13 00:04:19]

Can we inline this?

[Taylor Brandstetter, 2016/07/13 16:17:41]

Done.

 }
 
 }  // namespace cricket