Request for review (XS): 7200470 KeepAliveClosure not needed in CodeCache::do_unloading

Bengt Rutisson bengt.rutisson at
Tue Sep 25 05:51:50 PDT 2012

Hi David,

On 2012-09-25 14:26, David Holmes wrote:
> On 25/09/2012 9:59 PM, Bengt Rutisson wrote:
>> Thanks, Stefan!
>> I need one more review from an OpenJDK Reviewer. Any takers? It's a very
>> small change and fast to review...
> Looks okay.

Thanks for looking at this!

> I'm curious as to why the "keep alive" is no longer needed.

Discussed this a bit with Stefan. In the pre-perm-gen-removeal world the 
CodeCache needed the keep alive closure to handle the CompiledICHolders. 
Tom changed these from being oops to being native metadata structures. 
That means that they no longer move around and don't need to have their 
references updated by a keep alive closure.

> Makes me wonder if the actual KeepAliveClosures still serve a purpose?

I am not sure I fully understand your question. We still need keep alive 
closure for things that can find new objects - like reference 
processing. Is that what you are asking?


> David
> -----
>> Bengt
>> On 2012-09-25 13:20, Stefan Karlsson wrote:
>>> Looks good.
>>> StefanK
>>> On 09/24/2012 02:05 PM, Bengt Rutisson wrote:
>>>> Hi all,
>>>> Could I have a couple of reviews for this small change to remove some
>>>> unused code?
>>>> We are passing around a closure but we never use it.
>>>> Posting this on hotspot-dev since it is unclear to me if this is a
>>>> GC, Compiler or Runtime related fix. I plan on pushing this through
>>>> hotspot-gc.
>>>> Thanks,
>>>> Bengt

More information about the hotspot-dev mailing list