RFR: 8256811: Delayed/missed jdwp class unloading events [v17]

[Serguei Spitsyn]

On Wed, 13 Jul 2022 18:46:07 GMT, Chris Plummer <cjplummer at openjdk.org> wrote:

> `Actually I'm going to amend my previous comment about concerns with synchronization of the test. I think once the SetEventNotificationMode() call returns (after passing JVMTI_DISABLE), the caller should be guaranteed that any call to the event handler that was in progress has completed. So the caller should be able to safely free the raw monitor that is used in the callback. I'm not sure if JVMTI makes this guarantee however.`

The following RN is related to this question:
https://bugs.openjdk.org/browse/JDK-8256558
It explicitly says: `SetNotificationMode can be used to explicitly flush ObjectFree events, if needed.`
This RN is related to the bug pointed out by @zhengyu123 above: [JDK-8212879](https://bugs.openjdk.org/browse/JDK-8212879)
However, I feel that we still need to test it heavily to make sure there are no surprises.

-------------

PR: https://git.openjdk.org/jdk/pull/9168