qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Kennedy <andrewinternatio...@gmail.com>
Subject Re: Review Request: QPID-3216 - suggested fix for the deadlock issue.
Date Thu, 21 Apr 2011 15:02:39 GMT
On 20 April 2011 23:14, rajith attapattu <rajith77@gmail.com> wrote:
> Suggested fix for https://issues.apache.org/jira/browse/QPID-3216

This is for QPID-3214

Also, the 0-8 code does exactly the same thing, that is throws
exceptions to the connection listener if the session receives them, so
I'm not sure we can just ignore delivery. The problem being, if a
session fails, the only thing that can get an exception is the
connection, if it occurs asynchronously. IAs an alternative, maybe we
can remove the lock and replace it with an AtomicReference containing
the current exception, and also move the acquisition of the failover
mutex until after we have signalled the connection with the exception.
I will check how this behaves with the C++ profiles tonight.

(And, more generally, I don't view these as regressions, but as
continuing issues with locking that have been present for some
time...)

Andrew.
--
-- andrew d kennedy ? edinburgh : +44 7582 293 255

---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project:      http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org


Mime
View raw message