8371949: Upcall exception handler crashes with bad oop

[shipilev]

I have a local reproducer for JDK-8360595, and I only now understood why it crashed badly, without producing a proper error. Because with checked oops, oop is really a wrapper class, not a pointer. We are actually passing unwrapped oopDesc* from Java/stub code. All of our other entry points declare oopDesc*, this one should do the same.

I also extended the platform filter for the test, so reproducer can work on my x86 machines, and it now crashes more reliably:

$  CONF=linux-x86_64-server-fastdebug make images test TEST=java/foreign/TestUpcallStress.java TEST_VM_OPTS="-XX:+UnlockExperimentalVMOptions -XX:+UseEpsilonGC -XX:-ExitOnOutOfMemoryError"
...
[12:37:47.635] test TestUpcallStress.testUpcallsStress(663, "f1_V_SD_DID", VOID, [STRUCT, DOUBLE], [DOUBLE, INT, DOUBLE]): success [25ms]
Uncaught exception:
java.lang.OutOfMemoryError: Java heap space
<<no stack trace available>>

Additional testing:

• Linux x86_64 server fastdebug, reproducer no longer cryptically

---

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

---

Progress

• Change must not contain extraneous whitespace
• Commit message must refer to an issue
• Change must be properly reviewed (2 reviews required, with at least 1 Reviewer, 1 Author)

Issue

• JDK-8371949: Upcall exception handler crashes with bad oop (Bug - P4)

Reviewers

• Jorn Vernee (@JornVernee - Reviewer)
• David Holmes (@dholmes-ora - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/31151/head:pull/31151
$ git checkout pull/31151

Update a local copy of the PR:
$ git checkout pull/31151
$ git pull https://git.openjdk.org/jdk.git pull/31151/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 31151

View PR using the GUI difftool:
$ git pr show -t 31151

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/31151.diff

Using Webrev

Link to Webrev Comment

[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 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:

8371949: Upcall exception handler crashes with bad oop

Reviewed-by: jvernee, dholmes

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 9 new commits pushed to the master branch:

• f59ee5a: 8248122: java launcher should launch normally if JavaFX modules not found
• 9be6e77: 8384381: Remove AppContext from java.awt.EventQueue implementation
• aa29f79: 8298098: Remove SCCS id's from hotspot tests
• ... and 6 more: https://git.openjdk.org/jdk/compare/81442e1db839bd890719f9c183c7559b7fde889a...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.

[openjdk]

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

• core-libs
• hotspot

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.

[openjdk]

The total number of required reviews for this PR has been set to 2 based on the presence of this label: hotspot. This can be overridden with the /reviewers command.

[mlbridge]

Webrevs

• 00: Full (e4f96a63)

[dholmes-ora]

We pass oop parameters to lots of C++ methods. ??

[shipilev]

We pass oop parameters to lots of C++ methods. ??

"Pass" is too strong of the word. The actual problem is declaring oop in C++ method signature that is called from Java/stubs. oop is a wrapper, so you need to declare oopDesc*. This is what we do everywhere else.

[JornVernee]

Thanks for finding and fixing the issue. I've started a CI job to check if the test passes there on the new platforms as well. (From what I remember it had some stability issues on certain platforms)

[JornVernee]

All good from my end.

[dholmes-ora]

The actual problem is declaring oop in C++ method signature that is called from Java/stubs. oop is a wrapper, so you need to declare oopDesc*. This is what we do everywhere else.

Interesting limitation that I was not aware of. Not sure how to spot the places where this is necessary.