8369944: Notification can be lost due to interrupt in Object.wait

[theaoqi]

This is a clean backport of JDK-8369944 from mainline. The original change removes the redundant _notified field from ObjectWaiter and uses node.TState exclusively to determine whether a thread was notified.

On LoongArch64, this fix addresses a low-probability crash: guarantee(node.TState != ObjectWaiter::TS_WAIT) failed: invariant.

The crash is possibly due to the ordering between TState and _notified not being guaranteed on weakly-ordered architectures. Applying this patch resolves the crash.

---

• I confirm that I make this contribution in accordance with the OpenJDK Interim AI Policy.

---

Progress

• JDK-8369944 needs maintainer approval
• Change must not contain extraneous whitespace
• Commit message must refer to an issue

Issue

• JDK-8369944: Notification can be lost due to interrupt in Object.wait (Bug - P4)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk25u-dev.git pull/703/head:pull/703
$ git checkout pull/703

Update a local copy of the PR:
$ git checkout pull/703
$ git pull https://git.openjdk.org/jdk25u-dev.git pull/703/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 703

View PR using the GUI difftool:
$ git pr show -t 703

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk25u-dev/pull/703.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back aoqi! 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]

❗ This change is not yet ready to be integrated.
See the Progress checklist in the description for automated requirements.

[openjdk]

This backport pull request has now been updated with issue from the original commit.

[openjdk]

⚠️ @theaoqi This change is now ready for you to apply for maintainer approval. This can be done directly in each associated issue or by using the /approval command.

[mlbridge]

Webrevs

• 00: Full (40c4f704)

[theaoqi]

/approval request Clean backport of JDK-8369944. The fix removes the redundant _notified field from ObjectWaiter and uses node.TState exclusively to determine notification state, fixing a bug due to unordered reads of independent volatile fields. JDK-8369944 has been in JDK 26 since GA (~5 months) with no reported regressions. Locally tested tier1/tier2/tier3 on x86_64 (release and fastdebug), no issues found. The risk is low: the fix simplifies the code by removing a redundant state field.

[openjdk]

@theaoqi
8369944: The approval request has been created successfully.

[GoeLin]

Hi @theaoqi
please give a reason why you want to backport this.
Also, which is the change that causes this? How long does the problem exist?

[theaoqi]

Hi @GoeLin ,

Thanks for the reply.

The reason I am requesting this backport is as follows. It started because this backport fixes a crash on the LoongArch platform. However, note that OpenJDK does not officially support LoongArch (JIT), and this crash occurs in our self-maintained repository. Still, we think this issue may also affect other weakly-ordered architectures, which is why I am requesting this backport. I cannot confirm whether other platforms can actually reproduce this issue since I lack the testing environment.

The problem first became observable after JDK-8343840 and its corresponding LoongArch port, so it has been visible for over a year now.