Support using threadStatus field in JVM native #21043

fengxue-IS · 2025-01-30T21:54:30Z

Update thread status changes to j.l.Thread fields so getState() API
can be called without needing to halt the target thread.

Rework of #20550

Co-dependent extension repo changes:
JDK21: ibmruntimes/openj9-openjdk-jdk21#246
JDK25: ibmruntimes/openj9-openjdk-jdk25#43
JDKnext: ibmruntimes/openj9-openjdk-jdk#927

Fixes: #17251

Co-authored-by: Nathan Henderson nathan.henderson@ibm.com
Signed-off-by: Jack Lu Jack.S.Lu@ibm.com

fengxue-IS · 2025-01-30T21:58:56Z

Personal test build on JDK 8 & 21

fengxue-IS · 2025-01-31T14:41:43Z

@babsingh can you please take a look

babsingh

Changes LGTM functionally. Minor nits.

For JDK21+, the j.l.Thread implementation is different. Have you opened the extension repo PRs? Will we need to coordinate the merge between this PR and the extension repo PRs?

Which JDK versions (11,17,21) and test suites (sanity.functional,sanity.openjdk) have you used to test these changes in your personal build?

jcl/src/java.base/share/classes/java/lang/Thread.java

runtime/vm/threadhelp.cpp

runtime/vm/threadpark.cpp

jcl/src/java.base/share/classes/java/lang/Thread.java

babsingh · 2025-01-31T17:00:55Z

Can you also squash the commits and link this PR to the following perf issue: #17251?

babsingh · 2025-01-31T17:04:27Z

@gacholio Can you also review these changes?

fengxue-IS · 2025-01-31T17:10:35Z

For JDK21+, the j.l.Thread implementation is different. Have you opened the extension repo PRs? Will we need to coordinate the merge between this PR and the extension repo PRs?

Yes, I've opened PR for 21 & JDK next, see PR referenced above. will also include the link in PR description.
And yes the PR's would require a coordinated merge with extension repo

Which JDK versions (11,17,21) and test suites (sanity.functional,sanity.openjdk) have you used to test these changes in your personal build?

Testing have been done for 8,11,21 with sanity.functional & sanity.openjdk on xlinux

jcl/src/java.base/share/classes/java/lang/Thread.java

gacholio

Looks good as far as it goes. We should attempt to transition to using the state field everywhere instead of attempting to derive it from various other states and bits (in a future PR).

gacholio · 2025-02-03T17:41:18Z

Please squash the commits.

jcl/src/java.base/share/classes/java/lang/Thread.java

fengxue-IS · 2025-02-04T16:38:30Z

I've created the extension change for JDK24, will test removing the sync block in JDK21+ locally to see if it is still needed after the started check have been moved before checking eetop

gacholio · 2025-02-04T17:03:48Z

Beware the possibility of out-of-order memory operations.

fengxue-IS · 2025-02-06T22:10:17Z

squashed commits, updated extension changes

babsingh · 2025-02-06T22:59:59Z

translateJ9VMThreadStateToThreadState feedback for the switch case statement and missing param[in] also needs to be applied to the extension repo PRs.

fengxue-IS · 2025-02-14T18:51:20Z

translateJ9VMThreadStateToThreadState feedback for the switch case statement and missing param[in] also needs to be applied to the extension repo PRs.

done

babsingh · 2025-02-14T19:37:40Z

jenkins test sanity.openjdk,sanity.functional plinux jdk11

babsingh · 2025-02-14T19:38:15Z

jenkins test sanity.openjdk,sanity.functional zlinux jdk21 depends ibmruntimes/openj9-openjdk-jdk21#246

babsingh · 2025-02-14T19:38:37Z

jenkins test sanity.openjdk,sanity.functional zlinux jdk24 depends ibmruntimes/openj9-openjdk-jdk24#18

babsingh · 2025-02-14T19:38:57Z

Jenkins compile win jdk17

fengxue-IS · 2025-02-14T20:50:05Z

windows failure due to U_64 / U_32 conversion, will update VM_VMHelpers::setThreadState to use U_32 with explicit cast to avoid warning on windows. waiting until the other platform passes before pushing change.

runtime/oti/VMHelpers.hpp

pshipton · 2025-09-10T13:08:44Z

@fengxue-IS what's happening with this?

pshipton · 2025-09-10T13:10:30Z

This is blocking ibmruntimes/openj9-openjdk-jdk#927 which is needed to fix https://github.ibm.com/runtimes/openj9-openjdk-jdk21-zos/issues/522 which is marked as must fix.

fengxue-IS · 2025-09-10T22:00:35Z

@fengxue-IS what's happening with this?

There is some test failures on JDK24+ due to the addition of yielding virtual threads with monitors where the monitors in transition may have an incorrect state, another case is pinned virtual thread doesn't behave correctly for wait/notify due to the use of native API which doesn't update the virtual thread status.

I've pushed my latest change which have passed the known failures in local build/test.
Running personal build to verify there isn't any regression: https://hyc-runtimes-jenkins.swg-devops.com/view/OpenJ9%20-%20Personal/job/Pipeline-Build-Test-Personal/29845/

tajila · 2025-09-11T18:12:28Z

jenkins test sanity alinux64 jdk25 depends ibmruntimes/openj9-openjdk-jdk25#43

runtime/vm/threadpark.cpp

runtime/vm/vmthread.cpp

keithc-ca · 2025-09-11T19:30:21Z

Please add a comment in j9nonbuilder.h explaining that the values of J9VMTHREAD_STATE_* must not change because they are hard-coded in Thread.java.

fengxue-IS · 2025-09-17T14:51:55Z

@keithc-ca can you please take another look

keithc-ca

Please either squash, or arrange that changes to any given file are in a single commit. For example, VMHelpers.hpp should have the final change in the first (or only) commit.

runtime/oti/j9nonbuilder.h

jcl/src/java.base/share/classes/java/lang/Thread.java

runtime/vm/threadhelp.cpp

- Update thread status changes to j.l.Thread fields so getState() API can be called without needing to halt the target thread. - Add support of virtual thread state for JDK24+ Co-authored-by: Nathan Henderson <nathan.henderson@ibm.com> Signed-off-by: Jack Lu <Jack.S.Lu@ibm.com>

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

keithc-ca · 2025-09-17T20:54:48Z

Testing underway in

keithc-ca · 2025-09-17T20:55:18Z

Jenkins test sanity alinux64 jdk17

fengxue-IS force-pushed the threadstate2 branch 2 times, most recently from 6528e33 to aa845d1 Compare January 30, 2025 21:57

This was referenced Jan 30, 2025

Update getState() API to use holder.threadStatus field ibmruntimes/openj9-openjdk-jdk21#246

Merged

Update getState() API to use holder.threadStatus field ibmruntimes/openj9-openjdk-jdk#927

Merged

fengxue-IS marked this pull request as ready for review January 31, 2025 14:41

babsingh reviewed Jan 31, 2025

View reviewed changes

gacholio reviewed Feb 3, 2025

View reviewed changes

jcl/src/java.base/share/classes/java/lang/Thread.java Show resolved Hide resolved

gacholio approved these changes Feb 3, 2025

View reviewed changes

babsingh reviewed Feb 3, 2025

View reviewed changes

jcl/src/java.base/share/classes/java/lang/Thread.java Show resolved Hide resolved

fengxue-IS mentioned this pull request Feb 4, 2025

Update getState() API to use holder.threadStatus field ibmruntimes/openj9-openjdk-jdk24#18

Closed

babsingh approved these changes Feb 5, 2025

View reviewed changes

fengxue-IS force-pushed the threadstate2 branch 2 times, most recently from cd71984 to 09b5f25 Compare February 6, 2025 22:07

keithc-ca requested changes Mar 31, 2025

View reviewed changes

runtime/oti/VMHelpers.hpp Outdated Show resolved Hide resolved

fengxue-IS force-pushed the threadstate2 branch 4 times, most recently from cf021b8 to 15e11f3 Compare September 3, 2025 18:58

fengxue-IS force-pushed the threadstate2 branch from ff284ef to 01fc333 Compare September 10, 2025 21:41

fengxue-IS mentioned this pull request Sep 10, 2025

Update getState() API to use holder.threadStatus field ibmruntimes/openj9-openjdk-jdk25#43

Merged

fengxue-IS marked this pull request as ready for review September 11, 2025 17:56

keithc-ca requested changes Sep 11, 2025

View reviewed changes

runtime/vm/threadpark.cpp Outdated Show resolved Hide resolved

runtime/vm/vmthread.cpp Outdated Show resolved Hide resolved

fengxue-IS force-pushed the threadstate2 branch 3 times, most recently from 30dcc5d to 3b3a885 Compare September 16, 2025 19:33

fengxue-IS changed the title ~~[WIP] Support using threadStatus field in JVM native~~ Support using threadStatus field in JVM native Sep 16, 2025

fengxue-IS force-pushed the threadstate2 branch from 3b3a885 to 386b6f5 Compare September 16, 2025 19:43

keithc-ca requested changes Sep 17, 2025

View reviewed changes

runtime/oti/j9nonbuilder.h Outdated Show resolved Hide resolved

runtime/oti/j9nonbuilder.h Outdated Show resolved Hide resolved

jcl/src/java.base/share/classes/java/lang/Thread.java Show resolved Hide resolved

fengxue-IS force-pushed the threadstate2 branch from 386b6f5 to 073c99c Compare September 17, 2025 17:57

keithc-ca reviewed Sep 17, 2025

View reviewed changes

runtime/vm/threadhelp.cpp Outdated Show resolved Hide resolved

fengxue-IS and others added 2 commits September 17, 2025 14:37

Add JCL support for Java 8-17

0a58be1

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

fengxue-IS force-pushed the threadstate2 branch from 073c99c to 0a58be1 Compare September 17, 2025 18:38

keithc-ca approved these changes Sep 18, 2025

View reviewed changes

keithc-ca merged commit 1ee805a into eclipse-openj9:master Sep 18, 2025
6 checks passed

pshipton mentioned this pull request Sep 22, 2025

OpenJDK java/util/concurrent/Executors/UnreferencedExecutor OutOfMemory #22665

Open

Support using threadStatus field in JVM native #21043

Support using threadStatus field in JVM native #21043

Uh oh!

Conversation

fengxue-IS commented Jan 30, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

fengxue-IS commented Jan 30, 2025

Uh oh!

fengxue-IS commented Jan 31, 2025

Uh oh!

babsingh left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

babsingh commented Jan 31, 2025

Uh oh!

babsingh commented Jan 31, 2025

Uh oh!

fengxue-IS commented Jan 31, 2025

Uh oh!

Uh oh!

gacholio left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gacholio commented Feb 3, 2025

Uh oh!

Uh oh!

fengxue-IS commented Feb 4, 2025

Uh oh!

gacholio commented Feb 4, 2025

Uh oh!

fengxue-IS commented Feb 6, 2025

Uh oh!

babsingh commented Feb 6, 2025

Uh oh!

fengxue-IS commented Feb 14, 2025

Uh oh!

babsingh commented Feb 14, 2025

Uh oh!

babsingh commented Feb 14, 2025

Uh oh!

babsingh commented Feb 14, 2025

Uh oh!

babsingh commented Feb 14, 2025

Uh oh!

fengxue-IS commented Feb 14, 2025

Uh oh!

Uh oh!

pshipton commented Sep 10, 2025

Uh oh!

pshipton commented Sep 10, 2025

Uh oh!

fengxue-IS commented Sep 10, 2025

Uh oh!

tajila commented Sep 11, 2025

Uh oh!

Uh oh!

Uh oh!

keithc-ca commented Sep 11, 2025

Uh oh!

fengxue-IS commented Sep 17, 2025

Uh oh!

keithc-ca left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

keithc-ca commented Sep 17, 2025

Uh oh!

keithc-ca commented Sep 17, 2025

Uh oh!

Uh oh!

fengxue-IS commented Jan 30, 2025 •

edited

Loading

gacholio left a comment •

edited

Loading