RFR: Bulk integration from shenandoah/jdk8

Roman Kennke rkennke at redhat.com
Mon Dec 9 14:34:47 UTC 2019


I forgot to mention: this has baked in shenandoah/jdk8 for a couple of
weeks now, gone through our CI many times, has been tested manually for
equally many times, and I believe some early adopters have also picked
it up and run with success. It also comes with a whole lot of new and
improved tests under gc/shenandoah.

Roman

> Hello all,
> 
> I'd like to propose integration of outstanding changes and backports
> from shenandoah/jdk8.
> 
> This time it is rather big. Most significantly, it includes LRB,
> elimination of forwarding pointer, the new x86_32 port, the Traversal
> GC, plus a whole lot of related bugfixes.
> 
> The list of changes:
> http://cr.openjdk.java.net/~rkennke/aarch64-shenandoah-integration-2019-12-09/changesets.txt
> 
> Shared-only webrev:
> http://cr.openjdk.java.net/~rkennke/aarch64-shenandoah-integration-2019-12-09/webrev-shared-only.00/
> 
> Full webrev:
> http://cr.openjdk.java.net/~rkennke/aarch64-shenandoah-integration-2019-12-09/webrev-all.00/
> 
> There are quite many shared-code changes involved in this. However, most
> of them actually revert/remove existing shared-code changes vs upstream
> jdk8u, so the net effect is a drastic improvement. We added a few new
> hooks (for the LRB), but we kept it inside the following pattern
> wherever possible:
> 
> #ifdef INCLUDE_ALL_GCS
> if (UseShenandoahGC) {...
> }
> #endif
> 
> Also, this change includes some renaming and moving-around of
> (Shenandoah) files to keep Shenandoah code in Shenandoah subdirectories,
> and follow the naming scheme of >=11 as much as possible.
> 
> Can I please get a review of this?
> 
> Thanks,
> Roman
> 



More information about the shenandoah-dev mailing list