8297168: Provide a bulk OopHandle release mechanism with the ServiceThread #11254

dholmes-ora · 2022-11-21T01:56:34Z

When a JavaThread terminates it has to release all OopHandles that it uses. This can't be done by the thread itself due to timing issues, so it is handed-off to the ServiceThread to do it - ref JDK-8244997.

Initially there was only one OopHandle to handle - that of the threadObj, but since Loom there are another 3 OopHandles to process. The existing logic does the hand-off one OopHandle at a time but that is a potential synchronization bottleneck because each hand-off acquires the ServiceLock, enqueues the OopHandle, issues a notify to (potentially) wakeup the ServiceThread and then releases the ServiceLock. This can lead to high contention on the ServiceLock and also bad scheduling interactions with the ServiceThread.

This PR contains two commits. The first simply changes the API so that we pass the target JavaThread so that all 4 OopHandles can be extracted in one call. The second commit looks at streamlining things further by consolidating into a single OopHandleList instance.

As ServiceThread is a subclass of JavaThread` I wasn't concerned about it having detailed knowledge of the JavaThread implementation.

Testing:

tiers 1-3
checked still no memory leak (ref JDK-8296463)

Thanks.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8297168: Provide a bulk OopHandle release mechanism with the ServiceThread

Reviewers

Robbin Ehn (@robehn - Reviewer)
Coleen Phillimore (@coleenp - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk pull/11254/head:pull/11254
$ git checkout pull/11254

Update a local copy of the PR:
$ git checkout pull/11254
$ git pull https://git.openjdk.org/jdk pull/11254/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 11254

View PR using the GUI difftool:
$ git pr show -t 11254

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/11254.diff

…ceThread API

bridgekeeper · 2022-11-21T01:57:39Z

👋 Welcome back dholmes! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2022-11-21T02:00:42Z

@dholmes-ora The following label will be automatically applied to this pull request:

hotspot-runtime

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2022-11-21T02:04:22Z

Webrevs

robehn

Looks fine!

src/hotspot/share/runtime/serviceThread.cpp

openjdk · 2022-11-21T07:48:21Z

@dholmes-ora This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8297168: Provide a bulk OopHandle release mechanism with the ServiceThread

Reviewed-by: rehn, coleenp

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 57 new commits pushed to the master branch:

0696854: 8297299: SequenceInputStream should not use Vector
f0e99c6: 8297301: Cleanup unused methods in JavaUtilJarAccess
392ac70: 8297211: Expensive fillInStackTrace operation in HttpURLConnection.getOutputStream0 when no content-length in response
5a45c25: 8297164: Update troff man pages and CheckManPageOptions.java
f12710e: 8288047: Accelerate Poly1305 on x86_64 using AVX512 instructions
cd6a203: 8297348: make CONF=xxx should match if xxx is an exact match
817e039: 8297352: configure should check pandoc version
15e2e28: 8297353: Regenerated checked-in html files with new pandoc
b366d17: 8294073: Performance improvement for message digest implementations
57f5cfd: 8296399: crlNumExtVal might be null inside X509CRLSelector::match
... and 47 more: https://git.openjdk.org/jdk/compare/7b3984cb5a08edb99a233c28331c00b25457d664...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

dholmes-ora · 2022-11-21T12:11:05Z

Thanks for the review @robehn .

coleenp

So the choice was moving OopHandleList to a header file, or telling ServiceThread about these fields.
Since nothing else calls add_oop_handle_release , I really would have picked the former, and moved the declaration of OopHandleList to javaThread.hpp, constructed the list and passed it to the service thread in javathread.cpp.

coleenp

I'm still not happy that ServiceThread knows all about JavaThread's fields that need to be cleared. But all of the OopHandleList code that's in serviceThread.cpp should be moved to javaThread.cpp. And two new functions added to JavaThread for ServiceThread to call (has_oop_handles_to_release and release_oop_handles).
I still think it's better since JavaThread is the only thing that has the necessity of deferred release.

dholmes-ora · 2022-11-21T21:23:30Z

Thanks for looking at this @coleenp. As I said I didn't think ServiceThread knowing the details was an issue because ServiceThread is-a JavaThread. But sure I can move everything to JavaThread instead if that is preferable.

coleenp

Looks good! Thank you.

src/hotspot/share/runtime/javaThread.hpp

dholmes-ora · 2022-11-22T02:41:41Z

Thanks for the review @coleenp !

dholmes-ora · 2022-11-22T21:55:21Z

Thanks again @robehn and @coleenp !

/integrate

openjdk · 2022-11-22T21:57:43Z

Going to push as commit a6c418e.
Since your change was applied there have been 67 commits pushed to the master branch:

4a544bb: 8297091: New langtools test jdk/javadoc/doclet/testValueTag/TestValueFormats.java fails on machines with unexpected number format
b6dddf4: 8239801: [macos] java/awt/Focus/UnaccessibleChoice/AccessibleChoiceTest.java fails
260e4dc: 8295011: EC point multiplication improvement for secp256r1
fb6c992: 8296957: One more cast in SAFE_SIZE_NEW_ARRAY2
ccc6e16: 8291067: macOS should use O_CLOEXEC instead of FD_CLOEXEC
0ac5b55: 8297349: Parallel: Use correct claim value for CLD oop iteration in PSScavengeCLDClosure
932bf35: 8297333: Parallel: Remove unused methods in PCIterateMarkAndPushClosure
42c2037: 8297382: Test fails to compile after JDK-8288047
6d6046b: 8252713: jtreg time out of CtrlASCII.java seems to hang the Xserver.
88957a7: 8297147: UnexpectedSourceImageSize test times out on slow machines when fastdebug is used
... and 57 more: https://git.openjdk.org/jdk/compare/7b3984cb5a08edb99a233c28331c00b25457d664...master

Your commit was automatically rebased without conflicts.

openjdk · 2022-11-22T21:58:07Z

@dholmes-ora Pushed as commit a6c418e.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

dcubed-ojdk

Thumbs up!

src/hotspot/share/runtime/javaThread.hpp

dcubed-ojdk · 2022-11-22T22:02:46Z

src/hotspot/share/runtime/javaThread.hpp

+ private:
+  // List of OopHandles to be released - guarded by the Service_lock.
+  static OopHandleList* _oop_handle_list;
+  // Add our OopHandles to the list for the service thread to release.


nit - s/service thread/ServiceThread/

dcubed-ojdk · 2022-11-22T22:04:40Z

src/hotspot/share/runtime/javaThread.cpp

-  ServiceThread::add_oop_handle_release(_vthread);
-  ServiceThread::add_oop_handle_release(_jvmti_vthread);
-  ServiceThread::add_oop_handle_release(_extentLocalCache);
+  // Enqueue OopHandles for release by the service thread.


nit - s/service thread/ServiceThread/

dcubed-ojdk · 2022-11-22T22:05:23Z

src/hotspot/share/runtime/serviceThread.hpp

@@ -52,7 +52,6 @@ class ServiceThread : public JavaThread {

  // Add event to the service thread event queue.


nit - s/service thread/ServiceThread/
Not your typo, but since you're close by...

dcubed-ojdk · 2022-11-22T22:06:29Z

Sorry I didn't submit my review on-time...

dholmes-ora · 2022-11-22T22:08:48Z

Sorry I didn't submit my review on-time...

No problem Dan - thanks for taking a look. I think it is okay to talk about the "service thread" as a concept, just as we talk about "compiler threads" and "gc threads" without always using the classname ServiceThread.

dholmes-ora added 2 commits November 18, 2022 01:15

Version 1: simply enqueue all the OOpHandles in one call to the Servi…

6e42d25

…ceThread API

Version 2: combine all 4 OopHandles into one OopHandleList instance

76f558d

openjdk bot added the rfr Pull request is ready for review label Nov 21, 2022

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Nov 21, 2022

robehn approved these changes Nov 21, 2022

View reviewed changes

src/hotspot/share/runtime/serviceThread.cpp Outdated Show resolved Hide resolved

src/hotspot/share/runtime/serviceThread.cpp Outdated Show resolved Hide resolved

openjdk bot added the ready Pull request is ready to be integrated label Nov 21, 2022

Rehn comments

31d7318

coleenp reviewed Nov 21, 2022

View reviewed changes

coleenp suggested changes Nov 21, 2022

View reviewed changes

Move all the code to JavaThread - as requested by @coleenp.

9c43a62

coleenp approved these changes Nov 21, 2022

View reviewed changes

src/hotspot/share/runtime/javaThread.hpp Show resolved Hide resolved

dholmes-ora requested a review from robehn November 22, 2022 02:41

robehn approved these changes Nov 22, 2022

View reviewed changes

openjdk bot added the integrated Pull request has been integrated label Nov 22, 2022

openjdk bot closed this Nov 22, 2022

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Nov 22, 2022

dcubed-ojdk reviewed Nov 22, 2022

View reviewed changes

dholmes-ora deleted the 8297168-oopHandle-release branch June 27, 2023 06:15

8297168: Provide a bulk OopHandle release mechanism with the ServiceThread #11254

8297168: Provide a bulk OopHandle release mechanism with the ServiceThread #11254

dholmes-ora commented Nov 21, 2022 •

edited by openjdk bot

bridgekeeper bot commented Nov 21, 2022

openjdk bot commented Nov 21, 2022

mlbridge bot commented Nov 21, 2022 •

edited

robehn left a comment •

edited

openjdk bot commented Nov 21, 2022 •

edited

dholmes-ora commented Nov 21, 2022

coleenp left a comment

coleenp left a comment

dholmes-ora commented Nov 21, 2022

coleenp left a comment

dholmes-ora commented Nov 22, 2022

dholmes-ora commented Nov 22, 2022

openjdk bot commented Nov 22, 2022

openjdk bot commented Nov 22, 2022

dcubed-ojdk left a comment

dcubed-ojdk Nov 22, 2022

dcubed-ojdk Nov 22, 2022

dcubed-ojdk Nov 22, 2022

dcubed-ojdk commented Nov 22, 2022

dholmes-ora commented Nov 22, 2022

		@@ -52,7 +52,6 @@ class ServiceThread : public JavaThread {

		// Add event to the service thread event queue.

8297168: Provide a bulk OopHandle release mechanism with the ServiceThread #11254

8297168: Provide a bulk OopHandle release mechanism with the ServiceThread #11254

Conversation

dholmes-ora commented Nov 21, 2022 • edited by openjdk bot

Progress

Issue

Reviewers

Reviewing

bridgekeeper bot commented Nov 21, 2022

openjdk bot commented Nov 21, 2022

mlbridge bot commented Nov 21, 2022 • edited

Webrevs

robehn left a comment • edited

Choose a reason for hiding this comment

openjdk bot commented Nov 21, 2022 • edited

dholmes-ora commented Nov 21, 2022

coleenp left a comment

Choose a reason for hiding this comment

coleenp left a comment

Choose a reason for hiding this comment

dholmes-ora commented Nov 21, 2022

coleenp left a comment

Choose a reason for hiding this comment

dholmes-ora commented Nov 22, 2022

dholmes-ora commented Nov 22, 2022

openjdk bot commented Nov 22, 2022

openjdk bot commented Nov 22, 2022

dcubed-ojdk left a comment

Choose a reason for hiding this comment

dcubed-ojdk Nov 22, 2022

Choose a reason for hiding this comment

dcubed-ojdk Nov 22, 2022

Choose a reason for hiding this comment

dcubed-ojdk Nov 22, 2022

Choose a reason for hiding this comment

dcubed-ojdk commented Nov 22, 2022

dholmes-ora commented Nov 22, 2022

dholmes-ora commented Nov 21, 2022 •

edited by openjdk bot

mlbridge bot commented Nov 21, 2022 •

edited

robehn left a comment •

edited

openjdk bot commented Nov 21, 2022 •

edited