Request for review (s): 6814390 G1: remove the concept of non-generational G1
Y. Srinivas Ramakrishna
y.s.ramakrishna at oracle.com
Thu Aug 18 01:10:12 PDT 2011
Hi Bengt -- that should also make the method checkpointRootsInitial() dead, and
thus also record_concurrent_mark_init_start() [The clue is the guarantee
you deleted in the latter method at line 940 which would be violated in the case of G1Gen
because it would be tautologically false now.] I think the dead code contagion
does stop there, as far as i could tell from a browse of the code. Perhaps a good
IDE may find you a few more dead methods, who knows...
Rest looks good to me.
On 8/18/2011 12:28 AM, Bengt Rutisson wrote:
> Hi John,
> Thanks for the review!
>> Looks good to me. You may want to consider removing the CMCheckpointRootsInitialClosure - it's
>> only instantiated in the code you removed from concurrentMarkThread.cpp.
> Good point. I removed the CMCheckpointRootsInitialClosure class. Here is an updated webrev:
>> On 08/16/11 11:03, Bengt Rutisson wrote:
>>> Hi all,
>>> Could I have a couple of reviews for this change?
>>> G1 was originally designed to be able to run in a non-generational mode. This has not been used
>>> for a long time and no testing has been done. Thus, the code has bit rotted. We don't see the
>>> need for this feature anymore, so rather than fixing it we should remove it.
>>> This is actually a low priority item, but since it will make my next step (supporting young
>>> space sizing better) simpler I would like to get this CR out of the way.
>>> 6814390 G1: remove the concept of non-generational G1
More information about the hotspot-gc-dev