RFR (S): 8078669: G1 applies SurvivorAlignmentInBytes to both survivor and old gen

Thomas Schatzl thomas.schatzl at oracle.com
Mon May 18 09:01:05 UTC 2015


Hi Jon,

On Wed, 2015-05-13 at 14:15 -0700, Jon Masamitsu wrote:
> 
> On 5/13/2015 7:33 AM, Thomas Schatzl wrote:
> > Hi all,
> >
> >    can I have reviews for the following small fix: since JDK-8060025
> > survivor alignment has been applied to both survivor and old gen regions
> > (if enabled of course).
> >
> > This has gone unnoticed, because the tests somehow detected that in this
> > situation everyhing is fine due to "someone else that allocated
> > objects".
> >
> > The change fixes the problem (in g1Allocator.hpp), and removes this
> 
> The CR suggests that this change in g1Allocator.hpp is a workaround while
> waiting for a better fix.  Still true?

Initially I thought that this change would have noticable performance
impact. Perf testing showed that this is not the case. Since the
alternative I could come up with was basically replicating lots of code
using templates (one variant supporting survivor alignment, the other
not), I thought that since there is no perf impact, this solution would
be fine.

> 
> The change looks correct as a fix for the specific problem.
> 
> I'll let you discuss the tests with Dmitry.

Thanks,
  Thomas




More information about the hotspot-gc-dev mailing list