8292443: Weak CAS VarHandle/Unsafe tests should test always-failing cases

[shipilev]

I noticed that current VarHandle/Unsafe test cases do not test the always-failing cases. Strong CASes do test both always-success and always-failing scenarios, while weak CASes only test spurious-failure tests. We should be also testing always-failing cases for these. Weak CAS can spuriously fail, but it cannot spuriously succeed.

The PR looks large, but the actual changes are in *.template files, and the rest is generated from them.

Additional testing:

• Linux x86_64 fastdebug java/lang/invoke/VarHandles and compiler/unsafe tests
• Linux x86_32 fastdebug java/lang/invoke/VarHandles and compiler/unsafe tests
• Linux AArch64 fastdebug java/lang/invoke/VarHandles and compiler/unsafe tests

---

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-8292443: Weak CAS VarHandle/Unsafe tests should test always-failing cases

Reviewers

• Paul Sandoz (@PaulSandoz - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk pull/9892/head:pull/9892
$ git checkout pull/9892

Update a local copy of the PR:
$ git checkout pull/9892
$ git pull https://git.openjdk.org/jdk pull/9892/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 9892

View PR using the GUI difftool:
$ git pr show -t 9892

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/9892.diff

[bridgekeeper]

👋 Welcome back shade! 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]

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

• core-libs
• hotspot-compiler

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.

[mlbridge]

Webrevs

• 00: Full (dc6040b8)

[PaulSandoz]

Looks good.

[openjdk]

@shipilev 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:

8292443: Weak CAS VarHandle/Unsafe tests should test always-failing cases

Reviewed-by: psandoz

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 36 new commits pushed to the master branch:

• 494d387: 8280152: AArch64: Reuse runtime call trampolines
• 8b4e6ba: 8289332: Auto-generate ids for user-defined headings
• 0fc9263: 8291828: Reduce runtime of java.util.stream microbenchmarks
• e81210f: 8292352: 32-bit Windows build failures after JDK-8290059
• f45b840: 8292384: Convert AdapterHandlerTable to ResourceHashtable
• 0c67fba: 8289049: x86_32 build fails with GCC 11 due to newString646_US warning
• bf7d6d3: 7132413: [macosx] closed/javax/swing/AbstractButton/5049549/bug5049549.java fails on MacOS
• e230719: 8292448: Convert BitMapFragmentTable to ResourceHashtable
• f75da22: 8292395: [testbug] VectorGatherScatterTest.java fails on SVE with -XX:MaxVectorSize=8 after JDK-8288397
• 802ef38: 8292327: java.io.EOFException in InflaterInputStream after JDK-8281962
• ... and 26 more: https://git.openjdk.org/jdk/compare/e89abb7d2d902c6e6b3c9953fc6663de10db77b9...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.

[shipilev]

Still testing aarch64, but I expect it to complete fine. If there are no other comments, I'll integrate this soon.

[shipilev]

Finally, AArch64 testing is complete. Results look good. I am integrating to unblock other test stabilization work.

/integrate

[openjdk]

Going to push as commit e8bc879.
Since your change was applied there have been 36 commits pushed to the master branch:

• 494d387: 8280152: AArch64: Reuse runtime call trampolines
• 8b4e6ba: 8289332: Auto-generate ids for user-defined headings
• 0fc9263: 8291828: Reduce runtime of java.util.stream microbenchmarks
• e81210f: 8292352: 32-bit Windows build failures after JDK-8290059
• f45b840: 8292384: Convert AdapterHandlerTable to ResourceHashtable
• 0c67fba: 8289049: x86_32 build fails with GCC 11 due to newString646_US warning
• bf7d6d3: 7132413: [macosx] closed/javax/swing/AbstractButton/5049549/bug5049549.java fails on MacOS
• e230719: 8292448: Convert BitMapFragmentTable to ResourceHashtable
• f75da22: 8292395: [testbug] VectorGatherScatterTest.java fails on SVE with -XX:MaxVectorSize=8 after JDK-8288397
• 802ef38: 8292327: java.io.EOFException in InflaterInputStream after JDK-8281962
• ... and 26 more: https://git.openjdk.org/jdk/compare/e89abb7d2d902c6e6b3c9953fc6663de10db77b9...master

Your commit was automatically rebased without conflicts.

[openjdk]

@shipilev Pushed as commit e8bc879.

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

[dholmes-ora]

Committing unchanged files, other than a copyright year change, looks very odd here. Why was the SunMisUnsafe template updated but none of the tests actually changed?

[shipilev]

Committing unchanged files, other than a copyright year change, looks very odd here. Why was the SunMisUnsafe template updated but none of the tests actually changed?

X-UnsafeAccessTest.java.template generates both SunMisc* and JdkInternalMisc* files, using JdkInternalMisc condition to disambiguate the two. So, if we need to change JdkInternalMisc*, like this PR does, we need to change X-UnsafeAccessTest.java.template, which regenerates SunMisc* too. Since SunMisc* tests do not have weakCAS operations, only the copyright updates hit them.

You could, in principle, selectively commit only the "substantially affected" files, but I believe the rule for auto-generated files after template updates is to commit everything that was changed after the template change. This guarantees that regenerating files from the unchanged template yields no changes.

[dholmes-ora]

Ah! Sorry I thought there were two distinct template file here.