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

List:       openjdk-serviceability-dev
Subject:    Re: RFR (S) 8173658: JvmtiExport::post_class_unload() is broken for non-JavaThread initiators
From:       coleen.phillimore () oracle ! com
Date:       2019-11-25 13:52:36
Message-ID: 58811209-8eda-ab6f-19b4-08a98e01409e () oracle ! com
[Download RAW message or body]



On 11/25/19 12:52 AM, David Holmes wrote:
> Hi Coleen,
>
> On 23/11/2019 5:42 am, coleen.phillimore@oracle.com wrote:
>> Summary: call extension ClassUnload event as a deferred event from 
>> the ServiceThread and remove unsafe arguments
>
> Looks good.
>
> Minor nit:
>
> src/hotspot/share/prims/jvmtiExport.cpp
>
> assert(Thread::current()->is_Java_thread(), "must be called from 
> ServiceThread");
>
> we have an explicit check for is_service_thread available :)

Yes, thanks now I do.  Thanks for reviewing and your comments in the CSR.
Coleen

>
> Thanks,
> David
> -----
>
>> I'm still waiting for the CSR request to get approved but this change 
>> fixes the broken class unload events.  It's been tested with the 
>> existing test case, and hs-tier1 for all platforms and tier2-6 on 
>> linux-x64-debug.
>>
>> open webrev at 
>> http://cr.openjdk.java.net/~coleenp/2019/8173658.01/webrev
>> bug link https://bugs.openjdk.java.net/browse/JDK-8173658
>>
>> Thanks,
>> Coleen

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