RFR: Humongous object reclaim test

[Zhengyu Gu]

Hi Aleksey,

Thanks for the review.


On 09/21/2016 02:33 AM, Aleksey Shipilev wrote:
> On 09/20/2016 09:44 PM, Zhengyu Gu wrote:
>> Per Roman's request, this is the test case for humongous object reclaim
>> bug Roman pushed last week.
>> http://cr.openjdk.java.net/~zgu/humongous/webrev/
>   *) Copyright should mention Red Hat?

Fixed.

http://cr.openjdk.java.net/~zgu/humongous/webrev.01/

>
>   *) I wonder if the test should push the char[] reference on heap, and
> then carefully verify all its contents? Initializing the char[] off the
> pre-seeded Random object, and then verifying it with the same seed
> should confuse optimizers enough.

Given tiered compilation, I don't think it has chance to JIT. If I recall correctly,
the profiling counter has to reach 2000 to trigger C1.

Just to make sure, added VM flag "-Xint"


-Zhengyu


> Thanks,
> -Aleksey
>