RFR: 8271513: support JavaThreadIteratorWithHandle replacement by new ThreadsList::Iterator [v3]

[Coleen Phillimore]

On Thu, 5 Aug 2021 16:39:48 GMT, Daniel D. Daugherty <dcubed at openjdk.org> wrote:

>> src/hotspot/share/runtime/thread.cpp line 857:
>> 
>>> 855:     array[i] = _jvmci_old_thread_counters[i];
>>> 856:   }
>>> 857:   for (JavaThread* tp : ThreadsListHandle()) {
>> 
>> This is a nicer interface with underlying C++ magic.  There are lots of these in the source code.  Is it the intention to replace these as we go?
>
> There's definitely lots of these:
> 
> $ grep -r JavaThreadIteratorWithHandle src/hotspot | sed 's/:.*//' | sort | uniq -c
>    1 src/hotspot/os/linux/os_linux.cpp
>    1 src/hotspot/share/compiler/compileBroker.cpp
>    1 src/hotspot/share/gc/parallel/mutableNUMASpace.cpp
>    2 src/hotspot/share/gc/shared/collectedHeap.cpp
>    1 src/hotspot/share/gc/shared/gcLocker.cpp
>    1 src/hotspot/share/gc/shenandoah/shenandoahCodeRoots.cpp
>    5 src/hotspot/share/gc/shenandoah/shenandoahHeap.cpp
>    2 src/hotspot/share/gc/shenandoah/shenandoahPacer.cpp
>    1 src/hotspot/share/jfr/utilities/jfrThreadIterator.cpp
>    1 src/hotspot/share/jfr/utilities/jfrThreadIterator.hpp
>    1 src/hotspot/share/opto/idealGraphPrinter.cpp
>    1 src/hotspot/share/prims/jvmtiEnvBase.cpp
>    1 src/hotspot/share/prims/jvmtiEventController.cpp
>    1 src/hotspot/share/prims/jvmtiImpl.cpp
>    1 src/hotspot/share/prims/jvmtiTagMap.cpp
>    1 src/hotspot/share/prims/whitebox.cpp
>    5 src/hotspot/share/runtime/escapeBarrier.cpp
>    2 src/hotspot/share/runtime/handshake.cpp
>    1 src/hotspot/share/runtime/os.cpp
>    8 src/hotspot/share/runtime/safepoint.cpp
>    1 src/hotspot/share/runtime/thread.cpp
>    3 src/hotspot/share/runtime/threadSMR.hpp
>    3 src/hotspot/share/runtime/vmOperations.cpp
>    1 src/hotspot/share/services/heapDumper.cpp
>    1 src/hotspot/share/services/management.cpp
>    1 src/hotspot/share/services/threadService.cpp
>    1 src/hotspot/share/utilities/globalCounter.cpp
> 
> 
> I believe the plan is to migrate to this newer, cleaner interface.
> Obviously not with this bug, but down the road...

Awesome, thanks!

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

PR: https://git.openjdk.java.net/jdk/pull/4948