RFR [11]: 2018-10-18, Bulk backports to sh/jdk11

[Roman Kennke]

It looks good to me. Zhengyu should also look at it.

Thanks!!

Roman

> http://cr.openjdk.java.net/~shade/shenandoah/backports/jdk11-20181018/webrev.01/
> 
> There were many unclean applies and specific whitespace changes, please take a careful look.
> 
> Changes include:
> 
>  [backport] Fix and enable suspendible workers by default
>  [backport] Fix up superfluous changes against upstream
>  [backport] Move SATB flushing machinery out of shared code
>  [backport] Readjust ShenandoahStrDedupStress test to avoid timeout
>  [backport] Remove idle-on-uncommit support
>  [backport] Assert that Shenandoah does not trip JDK-8211926
>  [backport] Added a few missed worker sessions for worker tasks
>  [backport] Remove unused SA ShenandoahHeapRegionSet.java
>  [backport] lucene fixes
>  [backport] Add JFR parallel and concurrent events (infrastructure)
>  [backport] Remove ShenandoahCompileCheck
>  [backport] Clean up Shenandoah configurations in shared tests
>  [backport] Remove leftover from oopsHierarchy.hpp
>  [backport] A few more missing object equals barriers
>  [backport] c2 verification pass fixes for CAS
>  [backport] Move HdrSeq and BinaryMagnitudeSeq into Shenandoah utilities
>  [backport] Integrate Shenandoah tests to upstream test groups
>  [backport] Enable NUMA by default
>  [backport] Force termination logic is inverted
>  [backport] Reinstate single-threaded reference processing
>  [backport] Don't attempt to leave STS in final-traversal, where we never enter STS
>  [backport] Need to step over GC barriers in Node::eqv_uncast()
>  [backport] Consistenly order STS scope outside of OOM scope to avoid deadlocks
>  [backport] Disable suspendible workers (again)
>  [backport] Fixed ShenandoahStrDedupStress test that causes negative index OOB exception
> 
> Testing: tier3_gc_shenandoah {fastdebug|release}
> 
> Thanks,
> -Aleksey
>