Project proposal: Remove the Permanent Generation
Jon Masamitsu
jon.masamitsu at oracle.com
Mon Nov 28 11:24:12 PST 2011
Florian,
Thanks for the reminder that some changes have already
been made to aid in the removal of perm gen.
If anonymous classes are supported, they will complicate
matters some but there are other classes that we
will need to treat specially so it hopefully won't be too
much additional code. Last I heard, anonymous classes
were dropped from 292.
Jon
On 11/28/11 10:14, Florian Weimer wrote:
> * Jon Masamitsu:
>
>> I am going to be proposing the permanent generation
>> removal as a OpenJDK project shortly. The project is
>> described in the JEP 122
>>
>> http://openjdk.java.net/jeps/122
>>
>> This is not the formal proposal of the project but rather
>> a chance to ask questions ahead of that proposal.
> Interned strings have already been moved out of the permanent
> generation:
>
> changeset: 2226:b099aaf51bf8
> user: jcoomes
> date: Tue Mar 22 13:36:33 2011 -0700
> summary: 6962931: move interned strings out of the perm gen
>
> This is in jdk7u2 (and perhaps earlier).
>
> How does the PermGen removal interact with anonymous classes (that is,
> classes whose metadata can be reclaimed individually, while the
> corresponding class loader still exists)? I'm not sure what happened to
> anonymous classes; originally, this was one of the features in JDK 7
> targeted at dynamic languages. But if they still need to be supported,
> then this might complicate managing meta-data stored in the C heap.
>
More information about the hotspot-dev
mailing list