Request for Review (M) - 8150994: UseParallelGC fails with UseDynamicNumberOfGCThreads with specjbb2005

Thomas Schatzl thomas.schatzl at oracle.com
Tue Apr 26 08:12:06 UTC 2016


Hi Jon,

On Tue, 2016-04-12 at 20:50 -0700, Jon Masamitsu wrote:
> The ParallelGC collector compacts by moving live objects from source
> regions to destination regions.  The destination regions are regions 
> that have
> been emptied of live objects and the source regions are regions that
> hold
> the original locations of the live objects.  When a source region is 
> emptied,
> it becomes a destination region.  At the start of the compaction the
> destination regions are distributed to per thread lists processed by
> the GC
> workers.   The original implementation had tasks that drain the
> initial
> destination lists and tasks that steal destination regions from other
> GC workers.  The draining tasks and stealing tasks shared the per
> thread
> lists of destination regions.  The bug was that the sharing was not
> synchronized correctly.
> 
> See [1] for more details.
> 
> The fix is to combine the draining tasks and stealing tasks so that
> sharing of the destination lists is not needed.
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8150994
> 
> [2] http://cr.openjdk.java.net/~jmasa/8150994/webrev.00/
> 

  I think this looks good. The unusual qualifier because I am not very
knowledgable about how parallel gc works in that detail.

Thanks,
  Thomas


More information about the hotspot-gc-dev mailing list