RFR: JDK-8151336: Root region scanning should be cancelled and disabled when the ConcurrentMarkThread::run_service() exits

Bengt Rutisson bengt.rutisson at oracle.com
Wed Mar 16 12:50:41 UTC 2016


Hi Thomas,

Thanks for looking at this!

On 2016-03-16 12:46, Thomas Schatzl wrote:
> Hi Bengt,
>
> On Tue, 2016-03-15 at 16:11 +0100, Bengt Rutisson wrote:
>> Hi again everyone,
>>
>> Here's an updated webrev:
>> http://cr.openjdk.java.net/~brutisso/8151336/webrev.01
>>
>> I removed the extra _has_terminated state and am instead using the
>> ConcurrentMarkThread::should_terminate() state to make sure we don't
>> start a new initial mark if the concurrent mark thread is about to
>> terminate or has already terminated.
>>
>> The _should_terminate state is set in ConcurrentGCThread::stop()
>> where
>> we first take the Terminator_lock with a safepoint check. So, we
>> should
>> be able to rely on this value inside of
>> do_collection_pause_at_safepoint().
>    I would prefer if the _cmThread->should_terminate() condition should
> be inside decide_on_conc_mark_initiation(). It seems to be a quite
> generic condition that should apply anywhere.
>
> Although decide_on_conc_mark_initiation() has only this caller, this
> may cause bugs in the future if keeping it there.

I discussed this offline with Thomas. We agree that the proper way of 
doing this is to add information to the G1CollectorState class to keep 
track of whether or not the concurrent mark thread is running.

However, doing that in a thread safe manner requires some more thinking 
and testing. So, instead I will push what was proposed in webrev.01 and 
I filed this RFE to keep track of the remaining work:

JDK-8151995: Move CM Thread execution state into G1CollectorState
https://bugs.openjdk.java.net/browse/JDK-8151995

Thanks,
Bengt


> Thanks,
>    Thomas



More information about the hotspot-gc-dev mailing list