RFR: 2019-09-24, bulk backports to sh/jdk11

[Aleksey Shipilev]

On 9/23/19 6:48 PM, Aleksey Shipilev wrote:
> Let me backport these JVMTI fixes to sh/jdk11:
>   [backport] 8231197: Shenandoah: JVMTI heap walking cleanup crashes with NULL forwardee
>   [backport] 8231198: Shenandoah: heap walking should visit all roots most of the time
>   [backport] 8231244: Shenandoah: all-roots heap walking misses some weak roots
>   [backport] 8231249: Shenandoah: GC retries are too aggressive for tests that expect OOME
> 
> Webrev:
>   https://cr.openjdk.java.net/~shade/shenandoah/backports/jdk11-20190924/webrev.01/
> 
> Except the last patch in the queue, other patches have been tested in jdk/jdk over the weekend with
> no detected trouble. I would wait for next nightlies to verify the last patch too, and I don't
> foresee problems.

The last patch seems to cause some troubles in humongous allocation tests, so I would take that off
this backporting batch, and push the other three. It fixes a test timeouts, so the absence of it
should not affect product reliability. Meanwhile, I'll follow up on those failures.

-- 
Thanks,
-Aleksey