RFR (11): Streamline CAS barriers, C2 ideal graph generation and matching

[Roman Kennke]

>> Recently, I've streamlined the C1 and backend parts of CAS. When doing
>> the LRB backport to 11, I noticed that should also harmonize C2 ideal
>> generation and matching for CASes:
>>
>> http://cr.openjdk.java.net/~rkennke/jdk11-streamline-cas-c2/webrev.00/
> 
> Looks fine to me. Roland might want to eyeball?
> 
> Hold it off until we clear and push shenandoah-jdk-11.0.4+5 tag in the evening?
> 

I will surely wait for Roland, which would be early next week I suppose.

Roman