Author: mgoulish
Date: Wed Mar 30 19:11:09 2011
New Revision: 1087047
URL: http://svn.apache.org/viewvc?rev=1087047&view=rev
Log:
qpid-3171
The registration of the codec happens on a different thread from the
use of the codec. It is possible for the registration to occur after
the first attempted use. In my testing, this happened 3% of the time
-- 165 times out of 5000 tests -- when using RDMA transport, and 0 times
out of 5000 when using TCP. Which is why we didn't notice it earlier.
We have a function that tells when we are ready to encode --
CyrusSecurityLayer::canEncode. But it does not check the validity of
the codec pointer before using it, so it cores in this situation.
I believe simply checking that pointer is probably the best solution.
Introducing that check caused the crash not to show up in 10,000
trials. There were also no hangs.
Modified:
qpid/trunk/qpid/cpp/src/qpid/sys/cyrus/CyrusSecurityLayer.cpp
Modified: qpid/trunk/qpid/cpp/src/qpid/sys/cyrus/CyrusSecurityLayer.cpp
URL:
http://svn.apache.org/viewvc/qpid/trunk/qpid/cpp/src/qpid/sys/cyrus/CyrusSecurityLayer.cpp?rev=1087047&r1=1087046&r2=1087047&view=diff
==============================================================================
--- qpid/trunk/qpid/cpp/src/qpid/sys/cyrus/CyrusSecurityLayer.cpp (original)
+++ qpid/trunk/qpid/cpp/src/qpid/sys/cyrus/CyrusSecurityLayer.cpp Wed Mar 30
19:11:09 2011
@@ -106,7 +106,7 @@ size_t CyrusSecurityLayer::encode(const
bool CyrusSecurityLayer::canEncode()
{
- return encrypted || codec->canEncode();
+ return codec && (encrypted || codec->canEncode());
}
void CyrusSecurityLayer::init(qpid::sys::Codec* c)
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:[email protected]
