RFR(S): 7086533: G1: assert(!_g1->is_obj_dead(obj)): We should not be preserving dead objs: g1CollectedHeap.cpp:3835

John Cuthbertson john.cuthbertson at oracle.com
Tue Sep 27 22:40:19 UTC 2011


Hi Everyone,

Can I have a couple of volunteers review the changes for this CR? The 
webrev can be found at: http://cr.openjdk.java.net/~johnc/7086533/webrev.0/

Summary:
This assert trips while handling an evacuation failure during a 
partially young GC that takes place during a marking cycle. It is caused 
by some objects not being marking during an evacuation failure in a 
partially young GC during a previous marking cycle. We have decided to 
stop doing partially young GCs during marking cycles thus preventing the 
conditions that give rise to the bug.

Testing: the failing nightly tests; the GC test suite, and jprt.

Thanks,

JohnC



More information about the hotspot-gc-dev mailing list