8317262: LockStack::contains(oop) fails "assert(t->is_Java_thread()) failed: incorrect cast to JavaThread"

[dholmes-ora]

Please review this simple fix to LockStack::is_owning_thread() that allows it to be called by a non-JavaThread. Thanks to @pchilano for the regression test.

Testing:

• new regression test
• runtime/handshake/MixedHandshakeWalkStackTest.java
• tiers 1-3 sanity

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-8317262: LockStack::contains(oop) fails "assert(t->is_Java_thread()) failed: incorrect cast to JavaThread" (Bug - P3)

Reviewers

• Thomas Stuefe (@tstuefe - Reviewer)
• Patricio Chilano Mateo (@pchilano - Reviewer)
• Roman Kennke (@rkennke - Reviewer)
• Hamlin Li (@Hamlin-Li - Reviewer)

Contributors

• Patricio Chilano Mateo <pchilanomate@openjdk.org>

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/16047/head:pull/16047
$ git checkout pull/16047

Update a local copy of the PR:
$ git checkout pull/16047
$ git pull https://git.openjdk.org/jdk.git pull/16047/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 16047

View PR using the GUI difftool:
$ git pr show -t 16047

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/16047.diff

Webrev

Link to Webrev Comment

[bridgekeeper]

👋 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.

[dholmes-ora]

/contributor add @pchilano

[openjdk]

@dholmes-ora
Contributor Patricio Chilano Mateo <pchilanomate@openjdk.org> successfully added.

[openjdk]

@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]

Webrevs

• 00: Full (ca3835a7)

[tstuefe]

Looks reasonable.

[openjdk]

@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:

8317262: LockStack::contains(oop) fails "assert(t->is_Java_thread()) failed: incorrect cast to JavaThread"

Co-authored-by: Patricio Chilano Mateo <pchilanomate@openjdk.org>
Reviewed-by: stuefe, pchilanomate, rkennke, mli

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 109 new commits pushed to the master branch:

• 56aa1e8: 8317683: Add JIT memory statistics
• f7d6d7a: 8266242: java/awt/GraphicsDevice/CheckDisplayModes.java failing on macOS 11 ARM
• e942f36: 8317535: Shenandoah: Remove unused code
• 7cb2e6d: 8317514: Ensure MemorySegment is initialized before touching NativeMemorySegmentImpl
• 9622de2: 8317372: Refactor some NumberFormat tests to use JUnit
• 72c4dcb: 8317970: Bump target macosx-x64 version to 11.00.00
• 32a60cf: 8317824: Beef up javadoc for base offset in var handles derived from layouts (mainline)
• b12c471: 8317837: Leftover FFM implementation-only changes
• 605c976: 8318039: GHA: Bump macOS and Xcode versions
• 6273ab9: 8317808: HTTP/2 stream cancelImpl may leave subscriber registered
• ... and 99 more: https://git.openjdk.org/jdk/compare/3105538de5569845547b40f243a994a95a84b48f...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]

Thanks for the review @tstuefe - and welcome back :)

[stefank]

FWIW, there's a similar static function in sychronizer.cpp:

// Can be called from non JavaThreads (e.g., VMThread) for FastHashCode
// calculations as part of JVM/TI tagging.
static bool is_lock_owned(Thread* thread, oop obj) {
  assert(LockingMode == LM_LIGHTWEIGHT, "only call this with new lightweight locking enabled");
  return thread->is_Java_thread() ? JavaThread::cast(thread)->lock_stack().contains(obj) : false;
}

There might be an opportunity to combine/reuse/deduplicate this implementation.

[pchilano]

Looks good.

Patricio

[dholmes-ora]

Thanks for the review @pchilano .

[rkennke]

Looks good! Thanks!

[dholmes-ora]

Thanks for looking at this @stefank !

FWIW, there's a similar static function in sychronizer.cpp

They are not really that similar. Both are interested only in JavaThread's but one checks the current thread and the other checks a target thread.

[dholmes-ora]

Thanks for the review @rkennke !

[stefank]

Thanks for looking at this @stefank !

FWIW, there's a similar static function in sychronizer.cpp

They are not really that similar. Both are interested only in JavaThread's but one checks the current thread and the other checks a target thread.

Still. There's something fishy going on when the call sites look like this:
is_lock_owned => LockStack::contains => is_owning_thread

and both is_lock_owned and is_owning_thread filters away non-JavaThreads.

IMHO, I think that the entire StackWatermark handling should be pulled out of LockStack. Code that inspects oops of other threads need to start StackWatermark processing of the target thread. The StackWatermark processing inside LockStack looks like a workaround because some handshake operation forgot to do so. The patch seems to solve the immediate failure is OK to push, but I'd like to have a follow-up discussion on why we have this StackWatermark processing here.

[dholmes-ora]

I can't comment on the stackwatermark processing aspect. Just one follow up: is_owning_thread may be leading you astray with the name (it did me!) - it doesn't refer to the ownership of the lock/monitor but the lockStack itself. It would be more clearly named is_lock_stack_of_current_thread().

Thank again for looking at this.

[dholmes-ora]

/integrate

[openjdk]

Going to push as commit 4ea1b99.
Since your change was applied there have been 111 commits pushed to the master branch:

• 01ea1ef: 8305971: NPE in JavacProcessingEnvironment for missing enum constructor body
• 1d54e73: 8318072: DowncallLinker does not acquire/release segments in interpreter
• 56aa1e8: 8317683: Add JIT memory statistics
• f7d6d7a: 8266242: java/awt/GraphicsDevice/CheckDisplayModes.java failing on macOS 11 ARM
• e942f36: 8317535: Shenandoah: Remove unused code
• 7cb2e6d: 8317514: Ensure MemorySegment is initialized before touching NativeMemorySegmentImpl
• 9622de2: 8317372: Refactor some NumberFormat tests to use JUnit
• 72c4dcb: 8317970: Bump target macosx-x64 version to 11.00.00
• 32a60cf: 8317824: Beef up javadoc for base offset in var handles derived from layouts (mainline)
• b12c471: 8317837: Leftover FFM implementation-only changes
• ... and 101 more: https://git.openjdk.org/jdk/compare/3105538de5569845547b40f243a994a95a84b48f...master

Your commit was automatically rebased without conflicts.

[openjdk]

@dholmes-ora Pushed as commit 4ea1b99.

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