review request (M) - 8030646: track cset membership in one place

[Thomas Schatzl]

Hi,

On Wed, 2013-12-18 at 13:37 +0100, Bengt Rutisson wrote:
[...]
> 8030651-cset-unify-obj-in-cs
> 
> This is a nice cleanup. Thanks for fixing this!
> 
> I like the division between in_cset() and oop_in_cset(). Have you 
> checked if there are more occurrences of in_cset() that could be 
> replaced by oop_in_cset()? It looks to me like these in_cset() could be 
> changed to oop_in_cset() in these two methods:
> 
> G1CollectedHeap::handle_evacuation_failure_par()
> G1ParScanThreadState::verify_ref()
> 
> And I *think* G1STWIsAliveClosure::do_object_b() could also use 
> oop_in_cset() too.

  imo looking through all closures to fix obsolete/superfluous checks to
is_in_g1_reserved() or _g1_committed.contains() should be done in a
different CR systematically.

But I am not against doing so here.

At least for the G1ParCopyClosure there is already JDK-8027559.

Thomas