RFR(S): 8132721: Add tests which check that heap counters work as expected during Humongous allocations

Thomas Schatzl thomas.schatzl at oracle.com
Thu Feb 4 09:57:45 UTC 2016


Hi,

On Wed, 2016-02-03 at 14:30 -0800, Jon Masamitsu wrote:
> 
> 
> On 02/02/2016 02:09 PM, Kirill Zhaldbybin wrote:
> > Jon, 
> > 
> > Thank you for reviewing the fix. 
> > 
> > On 02/02/2016 10:53 PM, Jon Masamitsu wrote: 
> > >  
> > > Kirill, 
> > > 
> > > Would it make sense to add another full GC after line 177? 
> > I think you are right - it could make the test more stable. 
> > I also added additional output to separate this full GC from ones
> > that are called after deallocations. 
> > 
> > Here are a new webrev: 
> > http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8132721/webrev.06/


-  66     // Since during deallocation GC could free (very unlikely)
some non-humongous date this value relaxes amount of

s/date/data

- 176         System.out.println("Finished allocations - no GC should
happened before this line");

I think it should be "should have happened" instead of only happened,
but others may know better.

[...]
> Otherwise, looks good.  I don't need another
> webrev.
> 

 same here.

Thomas



More information about the hotspot-gc-dev mailing list