Openj9: JDK16 java/foreign/TestCleaner.java j9jcl.347 * ** ASSERTION FAILED ** at common/jdk_internal_misc_ScopedMemoryAccess.cpp:40: ((0 ))

[JasonFengJ9]

Failure link

From an internal build Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/286

22:14:14  openjdk version "16-internal" 2021-03-16
22:14:14  OpenJDK Runtime Environment (build 16-internal+0-adhoc.jenkins.BuildJDKnextppc64aixNightly)
22:14:14  Eclipse OpenJ9 VM (build master-9f11871a18a, JRE 16 AIX ppc64-64-Bit Compressed References 20201202_402 (JIT enabled, AOT enabled)
22:14:14  OpenJ9   - 9f11871a18a
22:14:14  OMR      - a58cd2f5687
22:14:14  JCL      - f7021059fd3 based on jdk-16+26)

To rebuild the failed tests in =https://hyc-runtimes-jenkins.swg-devops.com/job/Grinder, use the following links:
23:51:29 https://hyc-runtimes-jenkins.swg-devops.com/job/Grinder/parambuild/?JDK_VERSION=16&JDK_IMPL=openj9&BUILD_LIST=openjdk&PLATFORM=ppc64_aix&TARGET=jdk_foreign_0

Optional info

Failure output (captured from console output)

23:45:22  TEST: java/foreign/TestCleaner.java

23:45:22  STDERR:
23:45:22  WARNING: Using incubator modules: jdk.incubator.foreign
23:45:22  05:48:49.230 0x301c6800   j9jcl.347    *   ** ASSERTION FAILED ** at common/jdk_internal_misc_ScopedMemoryAccess.cpp:40: ((0 ))
23:45:22  JVMDUMP039I Processing dump event "traceassert", detail "" at 2020/12/02 23:48:49 - please wait.
23:45:22  JVMDUMP032I JVM requested System dump using '/home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/core.20201202.234849.11469214.0001.dmp' in response to an event
23:45:22  JVMDUMP010I System dump written to /home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/core.20201202.234849.11469214.0001.dmp
23:45:22  JVMDUMP032I JVM requested Java dump using '/home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/javacore.20201202.234849.11469214.0002.txt' in response to an event
23:45:22  JVMDUMP010I Java dump written to /home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/javacore.20201202.234849.11469214.0002.txt
23:45:22  JVMDUMP032I JVM requested Snap dump using '/home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/Snap.20201202.234849.11469214.0003.trc' in response to an event
23:45:22  JVMDUMP010I Snap dump written to /home/jenkins/workspace/Test_openjdknext_j9_sanity.openjdk_ppc64_aix_Nightly/openjdk-tests/TKG/test_output_1606969285655/jdk_foreign_0/work/scratch/Snap.20201202.234849.11469214.0003.trc
23:45:22  JVMDUMP013I Processed dump event "traceassert", detail "".

23:45:22  TEST RESULT: Failed. Unexpected exit from test [exit code: 255]
23:45:22  --------------------------------------------------

[pshipton]

@tajila fyi

The assert is Assert_JCL_unimplemented();

[tajila]

@babsingh Might be best to exclude this test in the short term until an investigation on how to properly implement this is determined. It may not be trivial to supply an appropriate implementation here as it deals with handling race conditions when closing scopes.

[babsingh]

Duplicate of https://github.com/eclipse/openj9/issues/11027#issuecomment-730849265, which reports the ScopedMemoryAccess failures:

Failed: java/foreign/TestSharedAccess.java
Failed: java/foreign/TestSpliterator.java
Failed: java/foreign/TestByteBuffer.java
Failed: java/foreign/TestHandshake.java
Failed: java/foreign/TestCleaner.java

may not be trivial to supply an appropriate implementation

This is true.

Might be best to exclude this test

The above failing tests are not blockers. Do we need to exclude these tests to stop generation of core files? Or can they continue to fail until they are fixed?

[pshipton]

Do we need to exclude these tests

Up to @JasonFengJ9, excluding the failures and getting green or nearly green makes it easier to triage the test results every day.

[pshipton]

Duplicate of #11027 (comment)

Do you want this issue open or closed?

[JasonFengJ9]

Do we need to exclude these tests
excluding the failures and getting green or nearly green makes it easier to triage the test results every day.

There are many other failures (such as those reported via https://github.com/eclipse/openj9/issues/11135) within sanity.openjdk, it is a bit distance from green.
Yes, the tests should be excluded.

Edit: created https://github.com/AdoptOpenJDK/openjdk-tests/pull/2085 to exclude the foreign memory access API tests.

[babsingh]

Do you want this issue open or closed?

+1 to close. I am tracking https://github.com/eclipse/openj9/issues/11027#issuecomment-730849265 to address all five test failures.

[tajila]

I would prefer to have separate issues for seperate features, ie foreign-memAccess/varhandles, sealedtypes, MH.

Makes it easier to assign things out.

[tajila]

Also makes it easier to determine what is fixed and what needs to be done. I see that some things in #11135 are already fixed.

@JasonFengJ9 is #11135 limited to just jdk.foreign test failures now?

[JasonFengJ9]

@tajila Besides jdk.foreign test failures (excluded via https://github.com/AdoptOpenJDK/openjdk-tests/pull/2085), sanity.openjdk (#11135) still have lots of java/lang/invoke/MethodHandles related failures which require OpenJDK JSR292 support (#7352) as per https://github.com/eclipse/openj9/issues/11135#issuecomment-724866324.
I am going to exclude them as well.

[pshipton]

@JasonFengJ9 pls help ensure we have an issue open for everything excluded. They need to be tagged with test excluded, which hopefully the committer will do, but pls double check.

[JasonFengJ9]

This test java/foreign/TestCleaner.java has been excluded via https://github.com/AdoptOpenJDK/openjdk-tests/pull/2085.