Daniel D. Daugherty wrote:
:
This fix adds the missing JVMTI GC callbacks for CMS:
Webrev: http://cr.openjdk.java.net/~kamg/6458402/webrev.03/
Very nicely done. You should ping Alan B. to see if he has
time to sanity check the JVM/TI tap map stuff...
I've just scanned it (haven't had time to do detailed review yet) but it
looks pretty good to me. The only thing I notice is one or two stale
comments. For example the hashmap() function no longer has an oop as
argument but is still commented as "return the appropriate hashmap for a
given object". I also wonder about the rehashing and whether it might be
cheaper to re-hash to same slot rather than accumulating the moved
entries in the delayed_add list. Keith may have measured this already.
-Alan.
