8268364: jmethod clearing should be done during unloading

[krk]

Backport of 8268364, single hunk conflict resolved by accepting "empty" from the backport commit vs. clear_jmethod_ids invocation in the dtor. Backport commit moves the clear_jmethod_ids invocation to unload method instead.

It fixes crashes when accessing jmethodIDs of a class being unloaded.

Here is the reproducer that crashes JVM in ~1 second without the patch, but works fine with it: gist.

This fix also resolves the issue reported at async-profiler/async-profiler#974 for Java 11.

To run the repro:

javac Main.java
gcc -shared -I"$JAVA_HOME/include" -I"$JAVA_HOME/include/linux" -fPIC repro.cpp -orepro.so

java -agentpath:"$(pwd)/repro.so" -Xmx100m -XX:+UseConcMarkSweepGC -XX:+CMSClassUnloadingEnabled -XX:+ExplicitGCInvokesConcurrent Main

---

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
• JDK-8268364 needs maintainer approval

Warning

 ⚠️ Found leading lowercase letter in issue title for 8268364: jmethod clearing should be done during unloading

Issue

• JDK-8268364: jmethod clearing should be done during unloading (Bug - P4 - Requested)

Reviewers

• Coleen Phillimore (@coleenp - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk11u-dev.git pull/2935/head:pull/2935
$ git checkout pull/2935

Update a local copy of the PR:
$ git checkout pull/2935
$ git pull https://git.openjdk.org/jdk11u-dev.git pull/2935/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 2935

View PR using the GUI difftool:
$ git pr show -t 2935

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk11u-dev/pull/2935.diff

Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back krk! 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.

[krk]

• MacOS builds are failing for unrelated reason: error: 'sprintf' is deprecated.
• linux-x86 / test - Test (tier1) failing while uploading test artifacts:

Error: Failed to CreateArtifact: Failed to make request after 5 attempts: Request timeout: /twirp/github.actions.results.api.v1.ArtifactService/CreateArtifact

[mlbridge]

Webrevs

• 01: Full - Incremental (032138d9)
• 00: Full (4cc13c48)

[krk]

cc @coleenp @fisk

[coleenp]

This looks like a good backport.

[openjdk]

⚠️ @krk 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.

[krk]

/approval request The fix is required to allow JVM TI tools, especially profilers, run safely with class unloading enabled. Low risk: a small localized change that only affects liveness of jmethodIDs.

[openjdk]

@krk
8268364: The approval request has been created successfully.

[krk]

/approval request All checks passing. The fix is required to allow JVM TI tools, especially profilers, to run safely with class unloading enabled. Low risk: a small localized change that only affects liveness of jmethodIDs.

[openjdk]

@krk
8268364: The approval request has been updated successfully.

[shipilev]

@GoeLin, @jerboaa -- do you see any blockers for approving this?

[jerboaa]

@shipilev JDK 11u is done and only very select fixes will be allowed in going forward. For this one, I'll let @theRealAph handle approval. On the surface it seems OK, but it's another change to a done release which only seems to affect corner cases (profilers when classes get unloaded).

FWIW, SAP stopped work on JDK 11u (11.0.25 was the last one) so you can not expect answers from them for 11u.