RFR(S): 8151637: Move CollectionSetChooser rebuild code into CollectionSetChooser (original) (raw)
Mikael Gerdin mikael.gerdin at oracle.com
Tue Mar 15 09:19:58 UTC 2016
- Previous message (by thread): RFR(S): 8151637: Move CollectionSetChooser rebuild code into CollectionSetChooser
- Next message (by thread): RFR(S): 8151637: Move CollectionSetChooser rebuild code into CollectionSetChooser
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi Thomas,
On 2016-03-15 09:36, Thomas Schatzl wrote:
Hi Mikael,
On Mon, 2016-03-14 at 12:07 +0100, Mikael Gerdin wrote: Hi all,
Currently when the CSetChooser needs to be injected with new regions the collector policy is responsible for calling a bunch of semi internal functions on the chooser and setting up some closures which are only relevant to the cset chooser. I've moved that code into the collection set chooser instead. I wasn't entirely clear on how the API for rebuild() should look since under the hood the closures go and access the G1CollectedHeap instance through the static in order to access the heap iteration method but since that has not changed in this patch I've left it as is. Bug: https://bugs.openjdk.java.net/browse/JDK-8151637 Webrev: http://cr.openjdk.java.net/~mgerdin/8151637/webrev.0/ Testing: JPRT looks good. Maybe the moved closure code (closure naming, g1 prefix, camelcase member names, make ParKnownGarbageHRClosure a local class to the task, etc) could be touched up a little. Feel free to ignore though.
In this case I'd prefer to leave it as is and move on to other changes. Thanks for the review.
/Mikael
Thanks, Thomas
- Previous message (by thread): RFR(S): 8151637: Move CollectionSetChooser rebuild code into CollectionSetChooser
- Next message (by thread): RFR(S): 8151637: Move CollectionSetChooser rebuild code into CollectionSetChooser
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]