RFR: ShenandoahControlThread should not get GCLABs

[Roman Kennke]

Am 31. August 2018 16:32:21 MESZ schrieb Aleksey Shipilev <shade at redhat.com>:
>On 08/31/2018 04:27 PM, Roman Kennke wrote:
>> Very good. Patch seems ok.
>> 
>> It will prevent the StringDedupThread from getting a GCLAB. Would
>that
>> be an issue?
>
>Yes, new RFR would deprive StringDedupThread from having a GCLAB.
>Which, I think, deserves a
>follow-up: why do we need to evacuate anything there (and in
>precleaning too?)

Well I guess that is the nature of traversal GC: anything that gets marked, e.g. via keep-alive closure, also gets evac'ed if it's in cset.

Roman