RFR (S) 8048215: [TESTBUG] java/lang/management/ManagementFactory/ThreadMXBeanProxy.java Expected non-null LockInfo

[David Holmes]

bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/

Simple race condition in the test. The main thread does checks that are
only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
easily reprodcued by just putting in a sleep after:

Object o = new Object();

Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread can't
proceed until wait() has been called and released the monitor.

Thanks,
David

[Daniel Fuchs]

Hi David,

Good finding! Looks reasonable to me.

best regards,

-- daniel

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/
Simple race condition in the test. The main thread does checks that are
only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread can't
proceed until wait() has been called and released the monitor.
Thanks,
David

[David Holmes]

Thanks for the review Daniel!

David

Post by Daniel Fuchs
Hi David,
Good finding! Looks reasonable to me.
best regards,
-- daniel

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/
Simple race condition in the test. The main thread does checks that
are only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread
can't proceed until wait() has been called and released the monitor.
Thanks,
David

[JC Beyler]

Hi David,

Looks good to me as well,
Jc

Post by David Holmes
Thanks for the review Daniel!
David

Post by Daniel Fuchs
Hi David,
Good finding! Looks reasonable to me.
best regards,
-- daniel

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/
Simple race condition in the test. The main thread does checks that
are only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread
can't proceed until wait() has been called and released the monitor.
Thanks,
David

[David Holmes]

Thanks Jc!

David

Post by Daniel Fuchs
Hi David,
Looks good to me as well,
Jc
Thanks for the review Daniel!
David

Post by Daniel Fuchs
Hi David,
Good finding! Looks reasonable to me.
best regards,
-- daniel

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/

<http://cr.openjdk.java.net/%7Edholmes/8048215/webrev/>

Post by Daniel Fuchs

Post by David Holmes
Simple race condition in the test. The main thread does checks that
are only valid once the target thread has called o.wait() but

there's

Post by Daniel Fuchs

Post by David Holmes
nothing to ensure that point of execution is reached. The

failure is

Post by Daniel Fuchs

Post by David Holmes
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits

upon and

Post by Daniel Fuchs

Post by David Holmes
for which the main also synchronizes on such that the main thread
can't proceed until wait() has been called and released the monitor.
Thanks,
David

--
Thanks,
Jc

[Mandy Chung]

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/
Simple race condition in the test. The main thread does checks that
are only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread
can't proceed until wait() has been called and released the monitor.

Thanks for fixing this.  It's good finds.  This looks good.

I wonder if some other ThreadMXBean tests have similiar bug.

Mandy

[David Holmes]

Thanks for the review Mandy!

David

Post by David Holmes
bug: https://bugs.openjdk.java.net/browse/JDK-8048215
webrev: http://cr.openjdk.java.net/~dholmes/8048215/webrev/
Simple race condition in the test. The main thread does checks that
are only valid once the target thread has called o.wait() but there's
nothing to ensure that point of execution is reached. The failure is
Object o = new Object();
Fix is to add a shared 'waiter' Object that the target waits upon and
for which the main also synchronizes on such that the main thread
can't proceed until wait() has been called and released the monitor.

Thanks for fixing this.  It's good finds.  This looks good.
I wonder if some other ThreadMXBean tests have similiar bug.
Mandy