Hi Volker,
Please note that non-trivial fixes should wait ~24hrs before integration
to ensure a range of folk have an opportunity to comment.
Thanks,
David
On 17/09/2021 5:59 am, Volker Simonis wrote:
On Thu, 16 Sep 2021 19:27:40 GMT, Coleen Phillimore <cole...@openjdk.org> wrote:
Currently, `OopHandle::release()` is implemented as follows:
inline void OopHandle::release(OopStorage* storage) {
if (peek() != NULL) {
// Clear the OopHandle first
NativeAccess<>::oop_store(_obj, (oop)NULL);
storage->release(_obj);
}
}
However, peek() returns NULL not only if the oop* `_obj` is NULL, but also when
`_obj` points to a zero oop. In the latter case, the oop* `_obj` will not be
released from the corresponding OopStorage and the slot it occupies will remain
alive forever.
This behavior can be easily triggered with the `LeakTestMinimal.java` test
which is attached to the [JBS
issue](https://bugs.openjdk.java.net/browse/JDK-8273902)(thanks to Oli
Gillespie from the Amazon Profiler team for detecting the issue and providing a
reproducer).
This fix should probably also be downported to jdk17 as quickly as possible.
Yes, please backport. Thank you for fixing this.
@coleenp, @sspitsyn thanks for the quick review!
-------------
PR: https://git.openjdk.java.net/jdk/pull/5549
