Open ben-walsh opened 5 years ago
Test excluded: https://github.com/AdoptOpenJDK/openjdk-tests/pull/630
Looking at this (might be same as https://github.com/eclipse/openj9/issues/3205)
Yeah, this was due to same cause as #3205, i.e, VarHandle
to a static field didn't initialize the field if the field in question has been initialized yet.
Manually verified that this test passes with the PR #3928 (already merged).
This issue can be closed (and the test can be re-enabled).
@smlambert Please re-enable the test. @ben-walsh would you kindly verify and close this issue?
Verification job running - https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/772 On successful completion, I'll create a PR to re-enable. @smlambert
https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/772 shows the problem is still present running against a nightly 11 build, which I assume is not expected ?
@JasonFengJ9 What was the build version you manually verified against ?
@ben-walsh I did verification with a patch generated from the PR.
Looked at the grinder job, its jvm -version
output is as following:
openjdk version "11.0.1" 2018-10-16
OpenJDK Runtime Environment AdoptOpenJDK (build 11.0.1+13-201901071931)
Eclipse OpenJ9 VM AdoptOpenJDK (build master-f8231edc8, JRE 11 Linux amd64-64-Bit Compressed References 20190107_86 (JIT enabled, AOT enabled)
OpenJ9 - f8231edc8
OMR - 3c42fa79
JCL - ab5eb2bc46 based on jdk-11.0.1+13)
https://github.com/eclipse/openj9/pull/3928 was merged on Jan. 8th so not contained in this jvm
.
Please re-run the grinder with latest nightly build (verified build 11.0.1+13-201901101911
passes this test).
Rebuilding https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/772 under https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/792. AOJDK website currently shows https://github.com/AdoptOpenJDK/openjdk11-binaries/releases/download/jdk11u-2019-01-11-01-22/OpenJDK11U-jdk_x64_linux_openj9_2019-01-11-01-22.tar.gz as the latest nightly build available for the platform, which is what I'd expect to be used.
Grinder/792 passes, console output shows 2019-01-11-01-22 is used:
17:43:40 curl -OLJks https://api.adoptopenjdk.net/v2/binary/nightly/openjdk11?openjdk_impl=openj9&os=linux&arch=x64&release=latest&type=jdk&heap_size=normal $ ssh-agent 17:43:49 unzip file: OpenJDK11U-jdk_x64_linux_openj9_2019-01-11-01-22.tar.gz ... 17:43:53 Run /home/jenkins/workspace/Grinder/openjdkbinary/j2sdk-image/bin/java -version 17:43:53 openjdk version "11.0.1" 2018-10-16 17:43:53 OpenJDK Runtime Environment AdoptOpenJDK (build 11.0.1+13-201901101911) 17:43:53 Eclipse OpenJ9 VM AdoptOpenJDK (build master-16eaa1c01, JRE 11 Linux amd64-64-Bit Compressed References 20190110_89 (JIT enabled, AOT enabled) 17:43:53 OpenJ9 - 16eaa1c01 17:43:53 OMR - 6830bfc1 17:43:53 JCL - ab5eb2bc46 based on jdk-11.0.1+13)
@ben-walsh this can be closed now, thanks.
Test failure at https://github.com/ibmruntimes/openj9-openjdk-jdk11/blob/openj9/test/jdk/java/util/concurrent/forkjoin/FJExceptionTableLeak.java#L103 ...
Digging deeper, execution of this ...
... yields ...
I don't see anything in the source code to explain why the exceptionTableLock variable set in the JCL ForkJoinTask impl is nullified ...
Log archive : FJExTabLeak.tar.gz
This test passes against Hotspot.