RFR [8]: Degenerated GC

[Roman Kennke]

Looks Ok

On Wed, Feb 28, 2018 at 12:10 PM, Aleksey Shipilev <shade at redhat.com> wrote:
> http://cr.openjdk.java.net/~shade/shenandoah/8u-degenerated-gc/webrev.01/
>
> I have finally managed to backport Degenerated GC to sh/jdk8u, and this was the only outstanding
> feature to backport. The failures I mentioned before were caused by special handling for
> SurrogateLockerThread that is 8u-only [1]. When that is resolved [2], the tests start to pass.
>
> Testing: hotspot_gc_shenandoah {fastdebug|release}, specjvm runs with verification
>
> Thanks,
> -Aleksey
>
> [1] http://hg.openjdk.java.net/shenandoah/jdk8u/hotspot/rev/8716f1b84d2f
> [2] Because SLT handling in handle_alloc_failure_evac drops evac_in_progress, this is the fix:
>
>   if (heap->is_evacuation_in_progress()) {
>     heap->entry_evac();
>
> -   if (check_cancellation_or_degen(ShenandoahHeap::_degenerated_evac)) return;
>   }
>
> + // Capture evacuation failures that might have happened during pre-evac in final mark.
> + // In this case, SLT handling would drop evac-in-progress, and we miss it if we are
> + // checking under the branch above.
> + if (check_cancellation_or_degen(ShenandoahHeap::_degenerated_evac)) return;
>