RFR (S): 8223736:jvmti/scenarios/contention/TC04/tc04t001/TestDescription.java fails due to wrong number of MonitorContendedEntered events

[Daniel D. Daugherty]

On 6/19/19 9:59 PM, serguei.spitsyn at oracle.com wrote:
> Sorry, forgot the  bug title to add to the email subject.
>
> Thanks,
> Serguei
>
> On 6/19/19 6:09 PM, serguei.spitsyn at oracle.com wrote:
>> Please review a fix for test bug:
>>   https://bugs.openjdk.java.net/browse/JDK-8223736
>>
>> Webrev:
>> http://cr.openjdk.java.net/~sspitsyn/webrevs/2019/8223736-mon-events-test.1/ 
>>

test/hotspot/jtreg/vmTestbase/nsk/jvmti/scenarios/contention/TC04/tc04t001.java
     L164:             if (enterEventsCount() == lastEnterEventsCount + 1) {
         I wonder if '==' should be '>=' to more resilient against
         multiple events. I tried to figure out the flow control
         for the event posting, but I couldn't figure it out with
         a quick look.

     L170:         if (enterEventsCount() != lastEnterEventsCount + 1) {
         Similarly I wonder if this should be:
                   if (enterEventsCount() == lastEnterEventsCount) {
         to simply detect that the count hasn't changed.

test/hotspot/jtreg/vmTestbase/nsk/jvmti/scenarios/contention/TC04/tc04t001/tc04t001.cpp
     No comments.

Thumbs up. My concerns above may not be correct depending on how the
test really works. Your call.

Dan



>>
>> Summary:
>>  It seems that waiting for 0.5 sec for a MonitorContendedEnter event 
>> in the
>>  increment() method sometime is not enough (especially when the JFR 
>> is enabled).
>>  The fix implement an approach to ensure the event has posted before 
>> the worker
>>  thread goes to the next iteration.
>>  Also, another check is added to diagnose if any of two worker threads
>>  (tc04t001Thread) has been interrupted by timeout.
>>  In fact, we have many other tests which miss this kind of check and 
>> diagnostics.
>>  We may want to consider fixing other cases if we encounter this 
>> eventually happens.
>>
>> Testing:
>>  A mach5 test submission is in progress.
>>
>> Thanks,
>> Serguei
>>
>