EVALUATION

ChangeSet=http://hg.openjdk.java.net/jdk7/hotspot-gc/hotspot/rev/7fbf850d87b7,ChangeRequest=6580131
                                     

EVALUATION

http://hg.openjdk.java.net/jdk6/jdk6/hotspot/rev/7fbf850d87b7
                                     

EVALUATION

http://hg.openjdk.java.net/jdk7/hotspot-rt/hotspot/rev/7fbf850d87b7
                                     

EVALUATION

ChangeSet=http://hg.openjdk.java.net/jdk7/tl/jdk/rev/b19cd193245e,ChangeRequest=6580131
                                     

EVALUATION

ChangeSet=http://hg.openjdk.java.net/jdk6/jdk6/jdk/rev/6073840fbd22,ChangeRequest=6580131
                                     

SUGGESTED FIX

See the attached 6580131-webrev-cr0.tgz file for the proposed fix.
This is the OpenJDK6 version of the fix.
                                     

EVALUATION

The folks at AMD Labs have been kind enough to provide an
extension to the JVM/TI CompileMethodLoad event in order
to provide additional information about in-lining. This
extension uses the existing (but previously unused)
compile_info paramter:

http://java.sun.com/javase/6/docs/platform/jvmti/jvmti.html#CompiledMethodLoad

Vasanth and company provided the HotSpot code changes and
the original demo program.

I'm listed as the RE for this bug, but I just did the Makefile
changes to export the new jvmticmlr.h file in the HotSpot repo
and the integration of the demo program into JAVA_HOME/demo/jvmti
in the JDK repo. I'm also handling the CCC request for getting
the jvmticmlr.h file added.