RFR: 8191333: Zero variant broken after 8189941

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Wed Nov 15 16:48:56 UTC 2017



On 11/15/17 11:04 AM, John Paul Adrian Glaubitz wrote:
> Hi Robbin!
>
> On 11/15/2017 04:56 PM, Robbin Ehn wrote:
>> Hi looks good, but:
>>
>> 101 #define SAFEPOINT       \
>>   102     if ( SafepointSynchronize::is_synchronizing()) {         \
>>   103         {         \
>>
>> You should not need to call SafepointSynchronize::is_synchronizing().
>> SafepointMechanism::block_if_requested(THREAD) takes care of everything.
> What about the handles?

Honestly, after looking at it, I have no idea.  It seems like something 
you should remove and see if that works.  The comment says that there 
should be no active handles.  Maybe have a NoHandleMark unconditionally? 
then SafepointMechanism::block_if_requested(THREAD) so that the comment 
is actually an assertion.  Sorry for too quick review.

Coleen
>
>> And if you ever plan to add thread-local handshake support to zero that
>> won't work.
> You mean, it won't work with the above code?
>
> Adrian
>



More information about the hotspot-dev mailing list