RFR (S): JDK-8058801: G1TraceReclaimDeadHumongousObjectsAtYoungGC only prints humongous object liveness output when there is at least one candidate humongous object
sangheon.kim at oracle.com
Thu Oct 9 17:29:40 UTC 2014
Thanks for the detail of this issue.
And thank you for sponsoring this too.
On 10/08/2014 04:01 AM, Thomas Schatzl wrote:
> Hi again,
> some clarification about the intent of this change:
> On Wed, 2014-10-08 at 12:58 +0200, Thomas Schatzl wrote:
>> On Tue, 2014-10-07 at 16:37 -0700, Sangheon Kim wrote:
>>> Hi all,
>>> can I have reviews for the following small change to print humongous
>>> objects information
>>> even though there's no candidates for humongous object?
>>> Main changes are:
>>> 1) Not to skip humongous reclaim phase if
>>> 'G1TraceReclaimDeadHumongousObjectsAtYoungGC' is set.
> The humongous reclaim phase has previously been skipped completely if
> there were no candidates in the beginning. This also skipped the trace
> messages enabled with G1TraceReclaimDeadHumongousObjectsAtYoungGC.
> However, with G1TraceReclaimDeadHumongousObjectsAtYoungGC enabled, it is
> still interesting to see information about the humongous objects that
> are in the system even if they cannot be reclaimed.
More information about the hotspot-gc-dev