RFR: 8233454: Test fails with assert(!is_init_completed(), "should only happen during init") after JDK-8229516

[David Holmes]

Hi Jie,

Thanks for filing this and attempting a fix. As per the bug report the 
underlying issue has now been fixed in Shenandoah, but I want to make 
the interrupt code more resilient as well:

http://cr.openjdk.java.net/~dholmes/8233454/webrev/

I was unable to reproduce the Shenandoah crash so if you could test this 
patch I would appreciate it - thanks. (Without the Shenandoah fix of 
course :) )

Meanwhile I'm putting the patch through other testing.

Thanks,
David
-----

On 4/11/2019 11:13 pm, David Holmes wrote:
> Hi Jie,
> 
> I will need to take a deeper look at this. This is a problem specific to 
> Shenadoah GC as it is triggering a sleep whilst a thread is still in the 
> process of attaching to the JVM :(
> 
> Thanks,
> David
> 
> On 4/11/2019 7:16 pm, Jie Fu wrote:
>> Hi all,
>>
>> JBS:    https://bugs.openjdk.java.net/browse/JDK-8233454
>> Webrev: http://cr.openjdk.java.net/~jiefu/8233454/webrev.00/
>>
>> According to the comment [1], the assert seems to miss the case for 
>> threads attached via JNI.
>> For more info, please refer to the JBS.
>>
>> Could you please review it and give me some advice?
>>
>> Thanks a lot.
>> Best regards,
>> Jie
>>
>> [1] 
>> http://hg.openjdk.java.net/jdk/jdk/file/2700c409ff10/src/hotspot/share/runtime/thread.hpp#l1249 
>>
>>
>>