8369013: Shenandoah: passive mode should support enabling ShenandoahCardBarrier #27966

rgithubli · 2025-10-24T00:18:03Z

Add card barriers to passive mode to test out the price of card barriers.

How this change is implemented is to instantiate the old region in passive mode - old region owns the card table so this would minimize the code change with a bit price of native memory. It does sound weird to have old gen in passive mode, but since passive mode is a just diagnostic mode, we'll go with it for the cleanliness of the change.

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-8369013: Shenandoah: passive mode should support enabling ShenandoahCardBarrier (Enhancement - P4)

Reviewers

William Kemper (@earthling-amzn - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/27966/head:pull/27966
$ git checkout pull/27966

Update a local copy of the PR:
$ git checkout pull/27966
$ git pull https://git.openjdk.org/jdk.git pull/27966/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 27966

View PR using the GUI difftool:
$ git pr show -t 27966

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/27966.diff

Using Webrev

Link to Webrev Comment

bridgekeeper · 2025-10-24T00:19:14Z

👋 Welcome back rgithubli! 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 · 2025-10-24T00:20:38Z

@rgithubli 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:

8369013: Shenandoah: passive mode should support enabling ShenandoahCardBarrier

Reviewed-by: wkemper

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 79 new commits pushed to the master branch:

d62553d: 8316274: javax/swing/ButtonGroup/TestButtonGroupFocusTraversal.java fails in Ubuntu 23.10 with Motif LAF
6080ccd: 8370797: Test runtime/ErrorHandling/AccessZeroNKlassHitsProtectionZone.java failed on macos 26
436dc68: 8367059: DTLS: loss of NewSessionTicket message results in handshake failure
... and 76 more: https://git.openjdk.org/jdk/compare/62f11cd4070f21ad82eebbb5319bdbbf4e13f9cf...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.

As you do not have Committer status in this project an existing Committer must agree to sponsor your change. Possible candidates are the reviewers of this PR (@earthling-amzn) but any other Committer may sponsor as well.

➡️ To flag this PR as ready for integration with the above commit message, type /integrate in a new comment. (Afterwards, your sponsor types /sponsor in a new comment to perform the integration).

openjdk · 2025-10-24T00:21:18Z

@rgithubli The following labels will be automatically applied to this pull request:

hotspot-gc
shenandoah

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

…ardBarrier

earthling-amzn

This turned out to be simpler than expected. Nice work. I suggest some small changes.

earthling-amzn · 2025-10-24T20:37:10Z

src/hotspot/share/gc/shenandoah/shenandoahHeap.cpp

+  // it means we're under passive mode and we have to initialize old gen
+  // for the purpose of having card table.
+  if (ShenandoahCardBarrier && !(mode()->is_generational())) {
+    _generation_sizer.heap_size_changed(max_capacity());


I think we could simplify here by not using the _generation_sizer. It should be fine to pass max_capacity() instead of max_capacity_old here.

Yeah that'd be simpler. Updating. Thanks.

earthling-amzn · 2025-10-24T20:38:02Z

src/hotspot/share/gc/shenandoah/shenandoahHeap.hpp

  }

  ShenandoahOldGeneration*   old_generation()    const {
-    assert(mode()->is_generational(), "Old generation requires generational mode");


Can we assert that ShenandoahCardBarrier is on instead of removing this assertion?

earthling-amzn · 2025-10-24T20:39:41Z

src/hotspot/share/gc/shenandoah/shenandoahHeap.hpp

  ConcurrentGCTimer*           _gc_timer;
  // For exporting to SA
  int                          _log_min_obj_alignment_in_bytes;
+  ShenandoahGenerationSizer     _generation_sizer;


I don't think we need this here.

earthling-amzn · 2025-10-24T20:44:01Z

src/hotspot/share/gc/shenandoah/shenandoahGenerationalHeap.hpp

  void initialize_heuristics() override;

  static ShenandoahGenerationalHeap* heap() {
-    shenandoah_assert_generational();


Can we put these assertions back? I don't see any code in the PR that would invalidate these assertions. I also don't see any changes that would instantiate the generational heap for non generational mode, so if there is code trying to use the generational heap, its behavior will be undefined.

It would crash at here with the assertion: code. Provided some of the stacktrace below. It's mostly because we enabled card barrier flag and the code just fell through and assumes it's generational.

Would it be more appropriate to change this line to be ShenandoahHeap* heap = ShenandoahHeap::heap();?

Stack: [0x00007f1082bd2000,0x00007f1082cd2000], sp=0x00007f1082cce370, free space=1008k Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code) V [libjvm.so+0x13e0e90] NativeStackPrinter::print_stack(outputStream*, char*, int, unsigned char*&, bool , int)+0x70 (shenandoahGenerationalHeap.hpp:47) V [libjvm.so+0x18fd2e9] VMError::report(outputStream*, bool)+0x1bff (vmError.cpp:979) V [libjvm.so+0x1900d01] VMError::report_and_die(int, char const*, char const*, __va_list_tag*, Thread*, unsigned char*, void const*, void const*, char const*, int, unsigned long)+0x8b1 (vmError.cpp:1887) V [libjvm.so+0xacf5ca] report_vm_status_error(char const*, int, char const*, int, char const*)+0x0 (deb ug.cpp:196) V [libjvm.so+0xacf33b] print_error_for_unit_test(char const*, char const*, __va_list_tag*)+0x0 (debug.c pp:149) V [libjvm.so+0x15bbec1] ShenandoahAsserts::assert_generational(char const*, int)+0x8b (shenandoahAssert s.cpp:532) V [libjvm.so+0x15be196] ShenandoahGenerationalHeap::heap()+0x19 (shenandoahGenerationalHeap.hpp:47) V [libjvm.so+0x16993bc] void card_mark_barrier<narrowOop>(narrowOop*, oopDesc*)+0x73 (shenandoahReferenceProcessor.cpp:65) V [libjvm.so+0x169a2ac] bool ShenandoahReferenceProcessor::discover<narrowOop>(oopDesc*, ReferenceType, unsigned int)+0x1da (shenandoahReferenceProcessor.cpp:409) V [libjvm.so+0x1698a5d] ShenandoahReferenceProcessor::discover_reference(oopDesc*, ReferenceType)+0xed (shenandoahReferenceProcessor.cpp:433) V [libjvm.so+0x1684aeb] bool InstanceRefKlass::try_discover<narrowOop, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0> >(oopDesc*, ReferenceType, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>*)+0x77 (instanceRefKlass.inline.hpp:74) V [libjvm.so+0x1682bfc] void InstanceRefKlass::oop_oop_iterate_discovery<narrowOop, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>, AlwaysContains>(oopDesc*, ReferenceType, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>*, AlwaysContains&)+0x2c (instanceRefKlass.inline.hpp:84) V [libjvm.so+0x1680acc] void InstanceRefKlass::oop_oop_iterate_ref_processing<narrowOop, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>, AlwaysContains>(oopDesc*, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>*, AlwaysContains&)+0x80 (instanceRefKlass.inline.hpp:111) V [libjvm.so+0x167ec2e] void InstanceRefKlass::oop_oop_iterate_ref_processing<narrowOop, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0> >(oopDesc*, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>*)+0x2c (instanceRefKlass.inline.hpp:134) V [libjvm.so+0x167ba88] void InstanceRefKlass::oop_oop_iterate<narrowOop, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0> >(oopDesc*, ShenandoahMarkRefsClosure<(ShenandoahGenerationType)0>*)+0x16c (instanceRefKlass.inline.hpp:154)

mlbridge · 2025-10-28T17:33:37Z

Webrevs

01: Full - Incremental (e6fdbc60)
00: Full (ecde5211)

earthling-amzn

Looks good. I think we should clear the card table for each degenerated cycle. Thank you for adding the test!

earthling-amzn · 2025-10-28T21:30:44Z

src/hotspot/share/gc/shenandoah/shenandoahFullGC.cpp

  }
+
+  // passive mode with ShenandoahCardBarrier turned on, clean the write table without swapping the tables
+  if (ShenandoahCardBarrier && !heap->mode()->is_generational()) {


In general, the passive mode will only initiate a degenerated cycle. It will run a full GC only if the degenerated cycle cannot evacuate or it doesn't make "good progress". Could we move this behavior into shDegeneratedGC or somewhere in shPassiveHeuristic? Cleaning the card table on every GC would be more aligned with the concurrent mode behavior.

earthling-amzn

Looks good. Thank you!

rgithubli · 2025-10-29T21:25:31Z

/integrate

openjdk · 2025-10-29T21:26:20Z

@rgithubli
Your change (at version e6fdbc6) is now ready to be sponsored by a Committer.

pengxiaolong · 2025-10-29T21:39:03Z

/sponsor

openjdk · 2025-10-29T21:40:40Z

Going to push as commit f3dfdfa.
Since your change was applied there have been 79 commits pushed to the master branch:

d62553d: 8316274: javax/swing/ButtonGroup/TestButtonGroupFocusTraversal.java fails in Ubuntu 23.10 with Motif LAF
6080ccd: 8370797: Test runtime/ErrorHandling/AccessZeroNKlassHitsProtectionZone.java failed on macos 26
436dc68: 8367059: DTLS: loss of NewSessionTicket message results in handshake failure
... and 76 more: https://git.openjdk.org/jdk/compare/62f11cd4070f21ad82eebbb5319bdbbf4e13f9cf...master

Your commit was automatically rebased without conflicts.

openjdk · 2025-10-29T21:40:49Z

@pengxiaolong @rgithubli Pushed as commit f3dfdfa.

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

openjdk bot added hotspot-gc [email protected] shenandoah [email protected] labels Oct 24, 2025

8369013: Shenandoah: passive mode should support enabling ShenandoahC…

3e11708

…ardBarrier

rgithubli force-pushed the 8369013-2 branch from 7724c67 to 3e11708 Compare October 24, 2025 00:22

earthling-amzn suggested changes Oct 24, 2025

View reviewed changes

rgithubli added 3 commits October 27, 2025 20:22

Addressed comments and added a test

7f0f807

Corrected copyright header

aec928d

Missed UnlockDiagnosticVMOptions

ecde521

rgithubli marked this pull request as ready for review October 28, 2025 17:29

openjdk bot added the rfr Pull request is ready for review label Oct 28, 2025

earthling-amzn reviewed Oct 28, 2025

View reviewed changes

Move write table clean to degen

e6fdbc6

earthling-amzn approved these changes Oct 29, 2025

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Oct 29, 2025

openjdk bot added the sponsor Pull request is ready to be sponsored label Oct 29, 2025

openjdk bot added the integrated Pull request has been integrated label Oct 29, 2025

openjdk bot closed this Oct 29, 2025

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review sponsor Pull request is ready to be sponsored labels Oct 29, 2025

8369013: Shenandoah: passive mode should support enabling ShenandoahCardBarrier #27966

8369013: Shenandoah: passive mode should support enabling ShenandoahCardBarrier #27966

Uh oh!

Conversation

rgithubli commented Oct 24, 2025 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issue

Reviewers

Reviewing

Uh oh!

bridgekeeper bot commented Oct 24, 2025

Uh oh!

openjdk bot commented Oct 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk bot commented Oct 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

earthling-amzn left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mlbridge bot commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

earthling-amzn left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

earthling-amzn left a comment

Choose a reason for hiding this comment

Uh oh!

rgithubli commented Oct 29, 2025

Uh oh!

openjdk bot commented Oct 29, 2025

Uh oh!

pengxiaolong commented Oct 29, 2025

Uh oh!

openjdk bot commented Oct 29, 2025

Uh oh!

openjdk bot commented Oct 29, 2025

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

3 participants

rgithubli commented Oct 24, 2025 •

edited by openjdk bot

Loading

openjdk bot commented Oct 24, 2025 •

edited

Loading

openjdk bot commented Oct 24, 2025 •

edited

Loading

mlbridge bot commented Oct 28, 2025 •

edited

Loading