Summary: Don't split if all edges will be moved to new phi
Reviewed-by: kvn

Reviewed-by: mdoerr, simonis
Contributed-by: NMatthew Brandyberry <mbrandy@linux.vnet.ibm.com>

Reviewed-by: ihse

Summary: Ignore error of intialization of C2 only variables
Reviewed-by: sla, dholmes

Reviewed-by: kvn

Summary: Set UseLoopCounter ergonomically if on-stack-replacement is enabled. Print warning.
Reviewed-by: kvn

Summary: CCP computes wrong type for CountedLoop iv Phi
Reviewed-by: kvn

Summary: Check for possible null-point dereference.
Reviewed-by: kvn

Reviewed-by: sla, dholmes

6651256: jstack: DeleteGlobalRef method call doesn't lead to descreasing of global refs count shown by jstack
Summary: jni_DeleteGlobalRef does not really release the jni handle, instead, set the handle point to JNIHandles::_deleted_handle which holds an oop instance (java/lang/Object) in Java heap and never be GC'ed. When counting number of global reference, it counts all the handles on the chain list, which includes the already deleted ones.
Reviewed-by: zgu, sla, coleenp
Contributed-by: yumin.qi@oracle.com

Summary: Bailout if OSR entry is unreachable.
Reviewed-by: thartmann
Contributed-by: NAndreas Woess <andreas.woess@oracle.com>

Reviewed-by: kvn, dholmes, bobv

Summary: The GetOwnedMonitorInfo() should not return a pending monitor
Reviewed-by: dcubed

Summary: Change TypeF:eq and TypeD:eq to compare NaN values using a bitwise comparison.
Reviewed-by: kvn
Contributed-by: NStefan Anzinger <stefan.anzinger@oracle.com>

Reviewed-by: alanb, sherman, ahgross, jeff