RFR 8217347 [TESTBUG] runtime/appcds/jvmti/InstrumentationTest.java timed out

David Holmes david.holmes at oracle.com
Mon Mar 25 04:14:11 UTC 2019

Hi Ioi,

On 25/03/2019 1:46 pm, Ioi Lam wrote:
> https://bugs.openjdk.java.net/browse/JDK-8217347
> http://cr.openjdk.java.net/~iklam/jdk13/8217347-appcds-InstrumentationTest-timeout.v01/ 
> This test case used VirtualMachine.list() to find the pid of the the child
> process. However, as David Holmes commented in the bug report, it's
> suspected that this call may lead to timeout.
> In any case, since Process.pid() has been added since JDK 9, it's better 
> to call this
> API directly. That way, we can avoid unnecessary dependency on 
> VirtualMachine.list().

In so far as I can see how you replaced the VM.list() with Process.pid() 
this seems okay. I'm having trouble actually understanding the handshake 
being used to control things though. I would have expected the target VM 
to create the flagfile to indicate it is ready for attaching, then it 
would wait for the flagfile to be deleted.


> Tested with tiers{1,2,3}. Also removed some unnecessary @module and 
> import lines.
> Thanks
> - Ioi

More information about the hotspot-runtime-dev mailing list