RFR: 8256811: Delayed/missed jdwp class unloading events [v18]

[Zhengyu Gu]

On Thu, 21 Jul 2022 13:33:01 GMT, Zhengyu Gu <zgu at openjdk.org> wrote:

>> Hi, it was added because some tests were failing.  With the concurrent posting, we were missing events and flushing them made the tests pass.
>
>> Hi, it was added because some tests were failing. With the concurrent posting, we were missing events and flushing them made the tests pass.
> 
> Thanks, @coleenp 
> Can you recall what were those tests? 
> 
> I removed the block and passed `vmTestbase_nsk_jdi`, `vmTestbase_nsk_jdwp` and `hotspot_serviceability` on Linux x86_64, I know Oracle has more tests.

> > @zhengyu123 try the tests with stress shendandoah flags if you have them. I think I had a lot of failures with ZGC and ZCG ZFragmentationLimit=0.1 or something like that.
> 
> [JDK-8255987](https://bugs.openjdk.org/browse/JDK-8255987)?

Tested on Linux x86_64 (fastdebug)
Test: `vmTestbase_nsk_jdi`, `vmTestbase_nsk_jdwp`, `vmTestbase_nsk_jvmti` `vmTestbase_nsk_jdb' and `jdk_jdi`

- [x] default, no failure
- [x] ShenandoahGC + aggressive, one timeout [1]
- [x] ZGC with -Xmx2g -XX:ZCollectionInterval=0.01, one timeout
- [x] ZGC with -XX:ZFragmentationLimit=0.1, no failure

[1] `vmTestbase/nsk/jvmti/scenarios/allocation/AP04/ap04t001/TestDescription.java`

I pushed update that removed the block, could you please run it through Oracle's internal tests?

Thanks!

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

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