RFR: Some trivial-ish cleanups

[Aleksey Shipilev]

On 06/11/2018 01:02 PM, Roman Kennke wrote:
> I swept over the current diff shenandoah vs upstream and fixed a number
> of trivial stuff that I found:
> - whitespaces here and there
> - reverted jniFastGetField_X.cpp stuff (we disabled this in Shenandoah
> for the time being. Current impl was buggy and found untrivial to fix).
> - MacroAssembler::in_heap_check() leftover from store-check
> - javaClasses.cpp revert leftover diff from arraycopy runtime barriers
> - barrierSet.hpp removed leftovers from assembly abstraction (moved to
> BarrierSetAssembler)
> - removed public read/write/keepalive barrier APIs. No longer needed
> - BarrierSetC1::_needs_null_check leftover from my C1 barriers merge.
> - accessBackend.cpp removed unneeded leftover (not present upstream)
> - unsafe.cpp removed remaining no longer needed barrier calls
> 
> http://cr.openjdk.java.net/~rkennke/cleanups/webrev.00/

Looks good!

I think that going forward we need to have ShenandoahBarrierSet::barrier_set() that would replace
odd casts like these:
   o = ((ShenandoahBarrierSet*) BarrierSet::barrier_set())->write_barrier(o);

Thanks,
-Aleksey