why LinkedBlockingQueue#poll(int, TimeUnit) is hanging up?
david.holmes at oracle.com
Tue Sep 11 02:05:30 UTC 2012
On 11/09/2012 11:43 AM, Vitaly Davidovich wrote:
> Symptoms here are eerily similar to
> http://bugs.sun.com/view_bug.do?bug_id=6822370, hence I'm curious if it
> still repros with -XX:+UseMembar.
Always an interesting question to ask but 6822370 was fixed in 7 and
this is reported in 7u5.
We really need a way to reproduce this, and see what all the other
threads are doing with respect to this queue.
> Sent from my phone
> On Sep 10, 2012 8:45 PM, "David Holmes" <david.holmes at oracle.com
> <mailto:david.holmes at oracle.com>> wrote:
> This kind of issue is better discussed, at least initially on Doug
> Lea's concurrency interest list:
> concurrency-interest at cs.__oswego.edu
> <mailto:concurrency-interest at cs.oswego.edu>
> The stack shows that it is the lock acquisition that is blocking
> indefinitely, not the poll itself. That is why it doesn't time out
> and why it hangs again immediately after you interrupt the thread.
> You need to see what is happening to other threads using this LBQ
> and whether one of them holds the lock and is itself "stuck"
> somewhere. You also need to see if there are any asynchronous
> exceptions involved (Thread.stop use, or StackOverflowException) as
> these can corrupt the internal state of an
> AbstractQueuedSynchronizer like ReentrantLock.
> On 11/09/2012 4:00 AM, Iris Clark wrote:
> Forwarding to core-libs-dev.
> -----Original Message-----
> From: Yegor Bugayenko [mailto:egor at technoparkcorp.__com
> <mailto:egor at technoparkcorp.com>]
> Sent: Monday, September 10, 2012 5:52 AM
> To: jdk7-dev at openjdk.java.net <mailto:jdk7-dev at openjdk.java.net>
> Subject: why LinkedBlockingQueue#poll(int, TimeUnit) is hanging up?
> I already posted the question here:
> Could one of you guys take a look? Thanks!
More information about the core-libs-dev