RFR: 8078021: SATB apply_closure_to_completed_buffer should have closure argument

Thomas Schatzl thomas.schatzl at oracle.com
Fri Apr 17 11:14:34 UTC 2015


Hi Kim,

On Fri, 2015-04-17 at 01:11 -0400, Kim Barrett wrote:
> Please review this simplification to SATB buffer processing.
> 
> SATBMarkQueueSet::apply_closure_to_completed_buffer() presently takes
> a worker_id index. It uses that to look up the closure associated with
> the worker by a preceding call to set_closure.
> 
> There is only one caller of this apply function,
> CMTask::drain_satb_buffers. It creates the desired closure, associates
> the closure with its worker id, calls
> apply_closure_to_completed_buffer, and clears the associated buffer.
> 
> We're changing the apply function to now take the closure directly, as
> an argument, replacing the worker_id argument that is now unused. We
> then eliminate set_closure and the supporting infrastructure, which
> are no longer used.
> 
> CR:
> https://bugs.openjdk.java.net/browse/JDK-8078021
> 
> Webrev:
> http://cr.openjdk.java.net/~kbarrett/8078021/webrev.00/
> 

  looks good.

Thomas




More information about the hotspot-gc-dev mailing list