RFR: Hook up Shenandoah barrier expand to ShBSC2::expand_macro_nodes()

[Roman Kennke]

I'm fine with that.

We don't strictly need to pass Compile* C if we already have
PhaseIterGVN& igvn, but that's ok either way for me.

Would you try to upstream the non-Shenandoah changes asap, and then push
the Shenandoah parts to shenandoah/jdk ?

Roman

>> Thoughts? Did I miss something?
> 
> I found pushing the barrier expand in a PhaseMacroExpand that we don't
> need quite ugly. What about this instead?
> 
> http://cr.openjdk.java.net/~roland/shenandoah/gcinterface-expand/webrev.00/
> 
> Roland.
>