RFR: JDK-8179259: Clean up BarrierSet

Erik Helin erik.helin at oracle.com
Tue Apr 25 12:48:23 UTC 2017

Hi Roman,

thanks for contributing! I don't have any comments about the patch but I 
might be able to answer your question about the JEP process, see my 
reply below.

On 04/25/2017 12:55 PM, Roman Kennke wrote:
> The BarrierSet class has lots of has_* and needs_* methods, none of
> which is used anywhere, except (some) in asserts (which is bogus too),
> and in one instance in a stub generator (which will be refactored to use
> the same paradigm as elsewhere). Those methods clutter the BarrierSet
> interface and need to be implemented in subclasses for no good reason.
> http://cr.openjdk.java.net/~rkennke/JDK-8163329/webrev.00/
> Tests: hotspot_gc, specjvm
> I also filed a small number of minor GC-interface related
> enhancement/refactoring bugs, all relatively isolated, all listed in the
> GC interface JEP. Will propose them soon, unless somebody stops me ;-)
> Oh, and maybe move the GC interface JEP to 'candidate' so that it gets a
> valid number?

I think Mark Reinhold is the one moving JEPs to the 'candidate' state 
since he is the OpenJDK Lead. Have a look at the "Workflow" section of 
the JEP 2.0 draft document:


> https://bugs.openjdk.java.net/browse/JDK-8163329
> Roman

More information about the hotspot-gc-dev mailing list