[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-serviceability-dev
Subject:    Re: RFR: 8282666: nsk/jvmti/PopFrame/popframe004 failed with: TEST FAILED: 30 JVMTI events were gene
From:       Serguei Spitsyn <sspitsyn () openjdk ! org>
Date:       2022-07-29 21:49:47
Message-ID: OluXWFCK_lFVFL3y-jSEb94U_5q_5OKkEPAlfF1fEVw=.71ad4fef-2161-426b-8683-efdeca12c87c () github ! com
[Download RAW message or body]

On Fri, 29 Jul 2022 19:08:35 GMT, Serguei Spitsyn <sspitsyn@openjdk.org> wrote:

> The nsk.jvmti test popframe004 is failing with the JVM args:
> `-Xcomp -XX:+CreateCoredumpOnCrash -ea -esa -XX:CompileThreshold=100 \
> -XX:+UnlockExperimentalVMOptions -XX:-TieredCompilation` 
> The call to join() below should not have a timeout parameter.
> 
> try {
> if (popFrameClsThr.isAlive())
> popFrameClsThr.join(2000);
> } catch (InterruptedException e) {
> out.println("TEST INCOMPLETE: caught " + e);
> totRes = FAILED;
> }
> if (popFrameClsThr.isAlive()) {
> out.println("TEST FAILED: thread with the method of " +
> "the popped frame is still alive");
> totRes = FAILED;
> } 
> 
> If the execution is slow by any reason (e.g. because of a compilation) then it \
> causes the main thread to fail joining with the target thread which is why the test \
> is failing with these symptoms. 
> This problem is hard to reproduce even in 1000's of runs on mach5.

Alex and Leonid, thank you for quick review!
I will integrate this now as the fix is pretty trivial.

-------------

PR: https://git.openjdk.org/jdk/pull/9689


[prev in list] [next in list] [prev in thread] [next in thread]