Trigger contended object monitor enter event for virtual threads by hangshao0 · Pull Request #21585 · eclipse-openj9/openj9

hangshao0 · 2025-04-07T15:29:41Z

issue #21545

hangshao0 · 2025-04-07T15:30:38Z

Change for J9HOOK_VM_MONITOR_CONTENDED_ENTERED will be added in a separate PR.

fengxue-IS

The code logic looks good, but missing support for other JVMTI functions which can be used by the user callback.

If this is sufficient to fix the test, I'm happy with moving the remaining support work into a different PR.

fengxue-IS

lgtm

babsingh · 2025-04-07T17:46:57Z

jenkins test sanity amac jdk24

babsingh · 2025-04-07T17:47:12Z

jenkins compile zlinux jdk21

hangshao0 · 2025-04-07T18:23:43Z

@fengxue-IS Adding VM_VMAccess::setPublicFlags(_currentThread, J9_PUBLIC_FLAGS_THREAD_BLOCKED) is causing assertion failure at thrinfo.c:240:

openj9/runtime/util/thrinfo.c

Line 240 in fcdc919

Assert_VMUtil_true(targetThread->blockingEnterObject != NULL);

#23 0x00007ffff6f931f9 in getVMThreadStateHelper (targetThread=0x254c00, pLockObject=0x0, pRawLock=0x0, pLockOwner=0x0, pCount=0x0, includeRawMonitors=0)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/util/thrinfo.c:240
#24 0x00007ffff6f6b9c0 in getThreadStateHelper (vmThread=0x254c00, threadObject=0x7fffb23e0, currentThread=0x254c00)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHelpers.cpp:742
#25 getVirtualThreadState (currentThread=0x254c00, thread=0x2586a8) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHelpers.cpp:876
#26 0x00007ffff6f835ff in jvmtiGetThreadState (env=<optimized out>, thread=0x2586a8, thread_state_ptr=0x7fffcae1312c)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiThread.cpp:95
--Type <RET> for more, q to quit, c to continue without paging--
#27 0x00007ffff426ac0c in _jvmtiEnv::GetThreadState (thread_state_ptr=0x7fffcae1312c, thread=0x2586a8, this=0x7ffff00c5818) at support/modules_include/java.base/jvmti.h:1197
#28 print_thread_info (jvmti=jvmti@entry=0x7ffff00c5818, jni=jni@entry=0x254c00, thread_obj=thread_obj@entry=0x2586a8) at test/lib/jdk/test/lib/jvmti/jvmti_common.hpp:258
#29 0x00007ffff426ad98 in MonitorContendedEnter (jvmti=0x7ffff00c5818, jni=0x254c00, thr=0x2586a8, obj=0x258670)
    at test/hotspot/jtreg/serviceability/jvmti/events/MonitorContendedEnter/mcontenter01/libmcontenter01.cpp:53
#30 0x00007ffff6f7077c in jvmtiHookMonitorContendedEnter (hook=<optimized out>, eventNum=<optimized out>, eventData=<optimized out>, userData=0x7ffff00c5818)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHook.c:2110
#31 0x00007ffff7a139de in J9HookDispatch (hookInterface=0x7ffff0011db0, taggedEventNum=<optimized out>, eventData=0x7fffcae135c0)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/util/hookable/hookable.cpp:237
#32 0x00007ffff7aed516 in VM_BytecodeInterpreterCompressed::yieldPinnedContinuation (returnState=<optimized out>, newThreadState=<optimized out>, _pc=<optimized out>, _sp=<optimized out>,
    this=<optimized out>) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:1554
#33 VM_BytecodeInterpreterCompressed::monitorenter (_pc=<optimized out>, _sp=<optimized out>, this=<optimized out>)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:8974
#34 VM_BytecodeInterpreterCompressed::run (this=0x7fffcae13840, vmThread=0x7ffff7471198) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:12012
#35 0x00007ffff7acfcde in bytecodeLoopCompressed (currentThread=<optimized out>) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.inc:112
#36 0x00007ffff7b9caf2 in c_cInterpreter () at /root/CCM/JDK24/openj9-openjdk-jdk24/build/linux-x86_64-server-release/vm/runtime/vm/xcinterp.s:157
#37 0x00007ffff7a36c12 in runJavaThread (currentThread=0x254c00) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/callin.cpp:682
#38 0x00007ffff7ab6098 in javaProtectedThreadProc (portLibrary=portLibrary@entry=0x7ffff7cc2420 <j9portLibrary>, entryarg=entryarg@entry=0x254c00)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/vmthread.cpp:2153
#39 0x00007ffff78c3451 in omrsig_protect (portLibrary=0x7ffff7cc2420 <j9portLibrary>, fn=0x7ffff7ab6000 <javaProtectedThreadProc(J9PortLibrary*, void*)>, fn_arg=0x254c00,
    handler=0x7ffff7a59f80 <structuredSignalHandler>, handler_arg=0x254c00, flags=506, result=0x7fffcae13d28) at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/port/unix/omrsignal.c:425
#40 0x00007ffff7ab232d in javaThreadProc (entryarg=0x7ffff000f9d0) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/vmthread.cpp:414
#41 0x00007ffff7a03c07 in thread_wrapper (arg=0x7ffff0dd9440) at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/thread/common/omrthread.c:1724

I see it is set to NULL at the end of preparePinnedVirtualThreadForUnmount().

openj9/runtime/vm/ContinuationHelpers.cpp

Line 1127 in fcdc919

J9VMTHREAD_SET_BLOCKINGENTEROBJECT(currentThread, currentThread, NULL);

fengxue-IS · 2025-04-07T20:01:14Z

@fengxue-IS Adding VM_VMAccess::setPublicFlags(_currentThread, J9_PUBLIC_FLAGS_THREAD_BLOCKED) is causing assertion failure at thrinfo.c:240:

openj9/runtime/util/thrinfo.c

Line 240 in fcdc919

Assert_VMUtil_true(targetThread->blockingEnterObject != NULL);

#23 0x00007ffff6f931f9 in getVMThreadStateHelper (targetThread=0x254c00, pLockObject=0x0, pRawLock=0x0, pLockOwner=0x0, pCount=0x0, includeRawMonitors=0)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/util/thrinfo.c:240
#24 0x00007ffff6f6b9c0 in getThreadStateHelper (vmThread=0x254c00, threadObject=0x7fffb23e0, currentThread=0x254c00)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHelpers.cpp:742
#25 getVirtualThreadState (currentThread=0x254c00, thread=0x2586a8) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHelpers.cpp:876
#26 0x00007ffff6f835ff in jvmtiGetThreadState (env=<optimized out>, thread=0x2586a8, thread_state_ptr=0x7fffcae1312c)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiThread.cpp:95
--Type <RET> for more, q to quit, c to continue without paging--
#27 0x00007ffff426ac0c in _jvmtiEnv::GetThreadState (thread_state_ptr=0x7fffcae1312c, thread=0x2586a8, this=0x7ffff00c5818) at support/modules_include/java.base/jvmti.h:1197
#28 print_thread_info (jvmti=jvmti@entry=0x7ffff00c5818, jni=jni@entry=0x254c00, thread_obj=thread_obj@entry=0x2586a8) at test/lib/jdk/test/lib/jvmti/jvmti_common.hpp:258
#29 0x00007ffff426ad98 in MonitorContendedEnter (jvmti=0x7ffff00c5818, jni=0x254c00, thr=0x2586a8, obj=0x258670)
    at test/hotspot/jtreg/serviceability/jvmti/events/MonitorContendedEnter/mcontenter01/libmcontenter01.cpp:53
#30 0x00007ffff6f7077c in jvmtiHookMonitorContendedEnter (hook=<optimized out>, eventNum=<optimized out>, eventData=<optimized out>, userData=0x7ffff00c5818)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/jvmti/jvmtiHook.c:2110
#31 0x00007ffff7a139de in J9HookDispatch (hookInterface=0x7ffff0011db0, taggedEventNum=<optimized out>, eventData=0x7fffcae135c0)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/util/hookable/hookable.cpp:237
#32 0x00007ffff7aed516 in VM_BytecodeInterpreterCompressed::yieldPinnedContinuation (returnState=<optimized out>, newThreadState=<optimized out>, _pc=<optimized out>, _sp=<optimized out>,
    this=<optimized out>) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:1554
#33 VM_BytecodeInterpreterCompressed::monitorenter (_pc=<optimized out>, _sp=<optimized out>, this=<optimized out>)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:8974
#34 VM_BytecodeInterpreterCompressed::run (this=0x7fffcae13840, vmThread=0x7ffff7471198) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.hpp:12012
#35 0x00007ffff7acfcde in bytecodeLoopCompressed (currentThread=<optimized out>) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/BytecodeInterpreter.inc:112
#36 0x00007ffff7b9caf2 in c_cInterpreter () at /root/CCM/JDK24/openj9-openjdk-jdk24/build/linux-x86_64-server-release/vm/runtime/vm/xcinterp.s:157
#37 0x00007ffff7a36c12 in runJavaThread (currentThread=0x254c00) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/callin.cpp:682
#38 0x00007ffff7ab6098 in javaProtectedThreadProc (portLibrary=portLibrary@entry=0x7ffff7cc2420 <j9portLibrary>, entryarg=entryarg@entry=0x254c00)
    at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/vmthread.cpp:2153
#39 0x00007ffff78c3451 in omrsig_protect (portLibrary=0x7ffff7cc2420 <j9portLibrary>, fn=0x7ffff7ab6000 <javaProtectedThreadProc(J9PortLibrary*, void*)>, fn_arg=0x254c00,
    handler=0x7ffff7a59f80 <structuredSignalHandler>, handler_arg=0x254c00, flags=506, result=0x7fffcae13d28) at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/port/unix/omrsignal.c:425
#40 0x00007ffff7ab232d in javaThreadProc (entryarg=0x7ffff000f9d0) at /root/CCM/JDK24/openj9-openjdk-jdk24/openj9/runtime/vm/vmthread.cpp:414
#41 0x00007ffff7a03c07 in thread_wrapper (arg=0x7ffff0dd9440) at /root/CCM/JDK24/openj9-openjdk-jdk24/omr/thread/common/omrthread.c:1724

I see it is set to NULL at the end of preparePinnedVirtualThreadForUnmount().

openj9/runtime/vm/ContinuationHelpers.cpp

Line 1127 in fcdc919

J9VMTHREAD_SET_BLOCKINGENTEROBJECT(currentThread, currentThread, NULL);

Yes, which is why I mentioned the 2 jvmti function at the same time. Once the preparePinnedVirtualThreadForUnmount is finished, the blocking state is transferred from the J9VMThread to Continuation object, the blocking object is stored in Continuation.blocker instead since we need to disassociate the virtual thread with its carrier.
For this PR, I suggest just reverting to the original changeset if that is sufficient to fix the test and patch both the state flag and blocker object together with the JVMTI function support.

issue eclipse-openj9#21545 Signed-off-by: Hang Shao <hangshao@ca.ibm.com>

hangshao0 · 2025-04-08T13:48:59Z

Reverted to the original change. The test mcontenter01 passed with the change.

babsingh · 2025-04-08T14:23:05Z

jenkins test sanity amac jdk24

babsingh · 2025-04-08T14:23:12Z

jenkins compile zlinux jdk21

babsingh · 2025-04-08T17:51:07Z

@hangshao0 There are failures in the sanity.openjdk PR build. Could you do an initial triage to see if it's related to your changes? If it's not, please create a new issue to track it. Once that's done, I'll go ahead and merge this PR.

hangshao0 · 2025-04-08T18:13:10Z

The test failures are known issues: #21446 and #21352.

hangshao0 · 2025-04-08T18:26:11Z

PR for 0.52: #21595

hangshao0 requested review from babsingh and fengxue-IS April 7, 2025 15:30

hangshao0 changed the title ~~Trigger contened object monitor enter envent for virtual threads~~ Trigger contended object monitor enter envent for virtual threads Apr 7, 2025

hangshao0 force-pushed the master branch from c748682 to 122c41b Compare April 7, 2025 15:31

hangshao0 changed the title ~~Trigger contended object monitor enter envent for virtual threads~~ Trigger contended object monitor enter event for virtual threads Apr 7, 2025

hangshao0 force-pushed the master branch from 122c41b to 1eb7f65 Compare April 7, 2025 15:33

fengxue-IS reviewed Apr 7, 2025

View reviewed changes

Comment thread runtime/vm/BytecodeInterpreter.hpp

hangshao0 force-pushed the master branch from 1eb7f65 to 337f291 Compare April 7, 2025 16:10

babsingh reviewed Apr 7, 2025

View reviewed changes

Comment thread runtime/vm/BytecodeInterpreter.hpp Outdated

hangshao0 force-pushed the master branch 3 times, most recently from 3e35164 to 3669be8 Compare April 7, 2025 17:39

fengxue-IS approved these changes Apr 7, 2025

View reviewed changes

Trigger contended object monitor enter event for virtual threads

2698e65

issue eclipse-openj9#21545 Signed-off-by: Hang Shao <hangshao@ca.ibm.com>

hangshao0 force-pushed the master branch from 3669be8 to 2698e65 Compare April 8, 2025 13:32

babsingh approved these changes Apr 8, 2025

View reviewed changes

babsingh merged commit 876dde0 into eclipse-openj9:master Apr 8, 2025

hangshao0 mentioned this pull request Apr 8, 2025

(0.52) Trigger contended object monitor enter event for virtual threads #21595

Merged

hangshao0 mentioned this pull request Apr 9, 2025

JDK24 JVMTI serviceability MonitorContendedEnter mcontenter01 #21404

Closed

This was referenced Apr 9, 2025

Re-include tests mcontenter01 and mcontentered01 adoptium/aqa-tests#6167

Merged

(1.0.6) Re-include tests mcontenter01 and mcontentered01 adoptium/aqa-tests#6169

Merged

(1.0.7) Re-include tests mcontenter01 and mcontentered01 adoptium/aqa-tests#6171

Merged

This was referenced Apr 9, 2025

JDK24 JVMTI serviceability MonitorContendedEntered mcontentered01 #21403

Closed

JDK24 JVMTI serviceability GetOwnedMonitorInfoTest #21406

Closed

This was referenced Apr 10, 2025

Re-enable GetOwnedMonitorInfoTest adoptium/aqa-tests#6174

Merged

(1.0.7) Re-enable GetOwnedMonitorInfoTest adoptium/aqa-tests#6175

Merged

Uh oh!

Conversation

hangshao0 commented Apr 7, 2025

Uh oh!

hangshao0 commented Apr 7, 2025

Uh oh!

fengxue-IS left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

fengxue-IS left a comment

Choose a reason for hiding this comment

Uh oh!

babsingh commented Apr 7, 2025

Uh oh!

babsingh commented Apr 7, 2025

Uh oh!

hangshao0 commented Apr 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

fengxue-IS commented Apr 7, 2025

Uh oh!

hangshao0 commented Apr 8, 2025

Uh oh!

babsingh commented Apr 8, 2025

Uh oh!

babsingh commented Apr 8, 2025

Uh oh!

babsingh commented Apr 8, 2025

Uh oh!

hangshao0 commented Apr 8, 2025

Uh oh!

hangshao0 commented Apr 8, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

hangshao0 commented Apr 7, 2025 •

edited

Loading