RFR (S): 8153170: Card Live Data does not correctly handle eager reclaim

Thomas Schatzl thomas.schatzl at oracle.com
Thu Apr 14 17:38:40 UTC 2016


Hi,

On Thu, 2016-04-14 at 11:30 -0400, Derek White wrote:
> Hi Thomas,
> 
> On 4/14/16 6:08 AM, Thomas Schatzl wrote:
> > Hi Kim,
> > [...]
> > 
> > Thanks a lot for catching these.
> > 
> > New webrevs at
> > http://cr.openjdk.java.net/~tschatzl/8153170/webrev.0_to_1/ (diff)
> > http://cr.openjdk.java.net/~tschatzl/8153170/webrev.1/ (full)
> > 
> > Thanks,
> >   Thomas
> This looks good. A couple of minor items. 
> 
> g1CardLiveData.cpp:
> 
> - Line 488 vs. (496, 512).
> The comment at 488 doesn't match the code. The comment talks about
> the old (expected && !actual) condition. Is (!expected && actual)
> somehow an OK condition or a failure? Or is the change part of having
> more exact card liveness information? 
> 

We want to be exact. I will remove the old comments.

> heapRegion.hpp
>   - Copyright.
> 
> I don't need to see an updated webrev though.
>  - Derek

Thanks for the review,
  Thomas



More information about the hotspot-gc-dev mailing list