RFR: 8308033: The jcmd thread dump related tests should test virtual threads
Jaikiran Pai
jpai at openjdk.org
Tue Apr 30 11:31:03 UTC 2024
On Tue, 30 Apr 2024 09:54:40 GMT, Jaikiran Pai <jpai at openjdk.org> wrote:
> Can I please get a review of these test-only changes which proposes to remove the `test/jdk/sun/tools/jcmd/JcmdOutputEncodingTest.java` and `test/jdk/sun/tools/jstack/BasicJStackTest.java` tests from `ProblemList-Virtual.txt`?
>
> When jtreg was enhanced to allow running the tests from within a virtual (main) thread, some tests were problem listed since they either were failing at that time or the test code would require additional work to make it work when the current thread is a virtual thread. The `test/jdk/sun/tools/jcmd/JcmdOutputEncodingTest.java` and `test/jdk/sun/tools/jstack/BasicJStackTest.java` are 2 such threads which require special handling when the current thread is a virtual thread.
>
> `test/jdk/sun/tools/jcmd/JcmdOutputEncodingTest.java` has been updated to use a different command to dump stacktraces when the test runs in a virtual thread. I went back and looked at the original issue for which this test was introduced and based on that, using a different command to dump the stacktraces shouldn't impact what the test was originally intended to test.
>
> `test/jdk/sun/tools/jstack/BasicJStackTest.java` has been updated to be skipped when the current thread is the virtual thread. This is because the test exercises the `jstack` tool which doesn't print stacktraces of virtual threads and thus the test isn't applicable for virtual threads.
>
> I've run these tests with this change, both with platform threads and virtual threads in our CI and the tests complete without failures.
> 8308033 is used in problem lists: [test/hotspot/jtreg/ProblemList-Virtual.txt]
It looks like the same bug id is being used in multiple different areas for different tests. I'll put this PR on draft for now, till I review if the following tests that are linked against this bug id can be removed from the problem listing or fixed too:
serviceability/dcmd/thread/PrintConcurrentLocksTest.java 8308033 generic-all
serviceability/dcmd/thread/PrintTest.java 8308033 generic-all
serviceability/dcmd/thread/ThreadDumpToFileTest.java 8308033 generic-all
serviceability/tmtools/jstack/DaemonThreadTest.java 8308033 generic-all
-------------
PR Comment: https://git.openjdk.org/jdk/pull/19016#issuecomment-2085044069
More information about the serviceability-dev
mailing list