Open JasonFengJ9 opened 10 months ago
Cause of this failure is #18394, the code jvmtiHeapFollowRefs_getStackData
depends on iteratorData->referrer
to set to proper value in e->refInfo.jni_local.thread_id
. while the fix in #18394 fixed issue where internal pointer in referrer field may be exposed to user code, it also removed ways to trace back to the thread that hold the JNI local.
This is not a regression but legacy issue with OpenJ9's impl of reference chain walker. moving this issue out of 0.42. This should be fixed as part of #18416
@fengxue-IS there isn't much point moving it to 0.43, it has almost the same time frame as 0.42. Moving it to 0.44.
@fengxue-IS @babsingh these tests should be excluded so they are not failing in every build. serviceability/jvmti/vthread/FollowReferences/VThreadStackRefTest.java#default serviceability/jvmti/vthread/FollowReferences/VThreadStackRefTest.java#no-vmcontinuations
these tests should be excluded so they are not failing in every build.
Opened https://github.com/adoptium/aqa-tests/pull/4937 to exclude the test.
@fengxue-IS What's the update on https://github.com/eclipse-openj9/openj9/issues/18416? Since this is a legacy issue, it doesn't block the current release. Do you agree on moving this issue to the 0.48 release?
@fengxue-IS What's the update on #18416? Since this is a legacy issue, it doesn't block the current release. Do you agree on moving this issue to the 0.48 release?
There was a prototype completed #18864, but the initial testing was flaky, so kept deferring until the test becomes stable. This doesn't affect anything in the current release, I agree we can move it to 0.48
Failure link
From an internal build(
sles12x86-rtp-rt6-1
):Rerun in Grinder - Change TARGET to run only the failed test targets.
Optional info
Failure output (captured from console output)
50x internal grinder - all failed
JDK21 ppc64le_linux JDK21 x86-64_mac