RFR [8]: 2018-06-01, Bulk backports to sh/jdk8

[Aleksey Shipilev]

On 06/01/2018 01:33 PM, Aleksey Shipilev wrote:
> http://cr.openjdk.java.net/~shade/shenandoah/backports/jdk8u-20180601/webrev.01/
> 
> This backports recent work to sh/jdk10:
> 
>  [backport] Incorrect label for static heuristics
>  [backport] Rename "cancel_concgc" to "cancel_gc"
>  [backport] Verifier should dump raw memory around the problematic oops
>  [backport] Move heuristics from ShCollectorPolicy to ShHeap
>  [backport] Rework GC degradation on allocation failure
>  [backport] Rework ClassUnloading* flags handling
>  [backport] Check heap stability in C1 WBs
>  [backport] ClassUnloadingWithConcurrentMark should be opt-in with Shenandoah
>  [backport] More verbose profiling for phase 4 in mark-compact
>  [backport] Full GC always comes with liveness data
>  [backport] Recycle the regions only once
>  [backport] Rename and move ShenandoahPrepareForMarkClosure
>  [backport] Reclaim immediate garbage after mark-compact marking
>  [backport] Shortcut regions that are known not to be alive
>  [backport] Refactor and improve ShenandoahCodeRoots strategies
> 
> We are missing the backport that enables ShenandoahCodeRootsStyle=2 by default, because upstream 8u
> does not work reliably with it yet. I'll take a look if we can backport some of upstream patches to
> upstream 8u to make this work.

Fixed! Atomic::add<jlong> again, this time in style-2 code root iterator. I folded the fix into
ShenandoahCodeRoots refactoring, and so I bring the remaining changeset:

  [backport] Default to ShenandoahCodeRootsStyle = 2

Full webrev:
  http://cr.openjdk.java.net/~shade/shenandoah/backports/jdk8u-20180601/webrev.02/

Still passes hotspot_gc_shenandoah {fastdebug|release}. I'll run a few more tests meanwhile, and
would like to push it somewhere around the weekend.

-Aleksey