RFR (S): 8129573: CollectedHeap::fill_with_objects() needs to use multiple arrays in 32 bit mode too

Tony Printezis tprintezis at twitter.com
Wed Jun 24 15:21:03 UTC 2015


Thomas,

It looks good. Only question I have is whether the change that sets filler_array_max_size based on the G1 heap region size has already gone in or not (as I couldn’t find it).

Tony

On June 24, 2015 at 5:40:33 AM, Thomas Schatzl (thomas.schatzl at oracle.com) wrote:

Hi all,  

can I have reviews for this small bugfix that got exposed with latest  
changes in JDK-8042668?  

With that change, we cannot assume that only one object is required to  
fill a given memory range, as filler_array_max_size is actually 512k  
with G1 now on 32 bit systems (i.e. typically you get 1M regions).  

The change just removes the #ifdef __LP64 and fixes the comment.  

The problem itself actually only got exposed with the upcoming change  
JDK-8129558, where GCBasher acts as a good reproducable test case.  

CR:  
https://bugs.openjdk.java.net/browse/JDK-8129573  

Webrev:  
http://cr.openjdk.java.net/~tschatzl/8129573/webrev  

Testing:  
JPRT (running gcbasher + containing JDK-8129558)  

Thanks,  
Thomas  



-----

Tony Printezis | JVM/GC Engineer / VM Team | Twitter

@TonyPrintezis
tprintezis at twitter.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/attachments/20150624/6497da71/attachment.html>


More information about the hotspot-gc-dev mailing list