Update Object.wait to use Blocker.begin/end for Java 19+

[fengxue-IS]

Fix: #16198

Signed-off-by: Jack Lu Jack.S.Lu@ibm.com

[fengxue-IS]

Personal build https://hyc-runtimes-jenkins.swg-devops.com/view/OpenJ9%20-%20Personal/job/Pipeline-Build-Test-Personal/14803/

@tajila can you please take a look
@0xdaryl can you review the JIT changes please, I only found 1 instance where Object.wait is hardcoded in JIT

[gacholio]

I think it might be best to add waitImpl to the final method list and leave wait in there - these are methods that do not require vTable entries (so we would save one slot in every class). Please give it a try and see if it causes any problems.

[fengxue-IS]

I think it might be best to add waitImpl to the final method list and leave wait in there - these are methods that do not require vTable entries (so we would save one slot in every class). Please give it a try and see if it causes any problems.

I've modified final.c to also include waitImpl and increased the OBJECT_FINAL_COUNT.
Did I miss any place that have a separate list?

[gacholio]

Sorry, I misread the diff - you have done exactly what I was suggesting.

[gacholio]

Is this the approach hotspot took? I'm concerned that people know too much about wait being native and this may cause issues with instrumentation agents.

[fengxue-IS]

RI have a similar approach to add Object.wait0
Even before this change, RI uses native wait(long milli) as the base where OpenJ9 uses native wait(long milli, int nano) as base.

[gacholio]

RI uses native wait(long milli) as the base

Does this mean they don't support nanos (ignore it and forward to the above native) or are there multiple natives?

[fengxue-IS]

RI uses native wait(long milli) as the base

Does this mean they don't support nanos (ignore it and forward to the above native) or are there multiple natives?

They round nanos to millis

[gacholio]

They round nanos to millis

Perhaps we should do the same (not in this PR).

Given the existing differences between the implementations, I believe instrumentation agents are unlikely to be affected by this change.

[fengxue-IS]

Addressed review comments, local testing passed on xlinux, personal build in progress: https://hyc-runtimes-jenkins.swg-devops.com/view/OpenJ9%20-%20Personal/job/Pipeline-Build-Test-Personal/14854/

[gacholio]

jenkins test sanity zlinux jdk8,jdk19

[gacholio]

jenkins compile win jdk19