Christian Despres

TestHashTablePerformance_2 failure with JITServer: collisions have been handled too slow expected [true] but found [false]

11

Failure link ------------ https://hyc-runtimes-jenkins.swg-devops.com/job/Test_openjdk11_j9_sanity.functional_x86-64_linux_jit_Personal/1678/ Optional info ------------- - [100x grinder with JITServer](https://hyc-runtimes-jenkins.swg-devops.com/job/Grinder/41082/) - 11/100 failures - [100x grinder without JITServer](https://hyc-runtimes-jenkins.swg-devops.com/job/Grinder/41094/) - 0/100 failures Failure output (captured from console output) ---------------------------------------------...

Document option -XX:[+|-]JITServerAOTCacheIgnoreLocalSCC

3

**Issue or pull request number:** Introduced in: https://github.com/eclipse-openj9/openj9/pull/18301 Made the default in: https://github.com/eclipse-openj9/openj9/pull/19585 **Overview:** This was previously an undocumented experimental option in 0.44.0, but since it's now enabled by default...

Non-fatal assert in extended.functional with JITServer and AOT cache: Must have shared VM access

4

The assert occurred with a debug build on JDK21 xlinux, with a manually-started JITServer on the side, with the client requesting the JITServer AOT cache be used, in the component...

Ensure deserializer class invalidation is complete

2

When a class is redefined, the old, invalidated RAM methods may be held in the "oldClass" or the "newClass", depending on whether or not the class was redefined in-place. To...

Allow SVM assertions to be ignored

1

If the option TR_IgnoreAssert is enabled for a particular compilation and the environment variable TR_svmAssertionsAreFatal is not set, SVM assertions will now be non-fatal even in DEBUG or PROD_WITH_ASSUMES builds....

SVM assert triggered in jdk_lang_0 with JITServer AOT cache: compileRomClass and currentRomClass should not be different!

1

I observed this at a client JVM when running `jdk_lang_0` with a manually-started JITServer on the side, using `EXTRA_OPTIONS` to enable the JITServer AOT cache on the client. This is...

Non-fatal assert triggered in sanity.functional with JITServer and AOT cache during ROM class packing

1

I encountered this when running `sanity.functional` manually with JITServer+AOT cache (with an appropriate JITServer instance started up manually on the side) with a debug compiler configuration to test the code...

JDK11 s390x linux jit build failed due to "Illegal instruction exception"

2

Observed internally in the JITServer nightly test pipeline, in this build: https://hyc-runtimes-jenkins.swg-devops.com/job/Build_JDK11_s390x_linux_jit_Personal/1524/. Note that the compiler is not built with `-XX:+UseJITServer` as far as I am aware - only the...

Non-fatal assert triggered in jdk_lang_0: 64-bit displacement should have been replaced

8

I observed this when running `sanity.openjdk` locally with a custom debug build. This was with a manually-started JITServer on the side, with the client requesting the JITServer AOT cache be...

Clang informs me that in this block of code, responsible for relocating a freshly-compiled AOT method: https://github.com/eclipse-openj9/openj9/blob/110ef95462c09cd7b94e75500133166b671d12fe/runtime/compiler/control/CompilationThread.cpp#L10417-L10432 we create and then immediately destroy the `TR::CompilationInfoPerThreadBase::InterruptibleOperation` object that is created, because...