Fix for a missing allocationFence in transformArrayCloneCall()

When we transform a clone() of an array object into a newarray and
arraycopy, we neglected to add an allocationFence. This change will
introduce an allocationFence after the arraycopy to ensure that all
threads see the new contents of memory on weak memory coherency
machines like POWER and AArch64.

Signed-off-by: Kevin Langman <langman@ca.ibm.com>

compiler/optimizer/ValuePropagationCommon.cpp

@@ -4136,6 +4136,15 @@ void OMR::ValuePropagation::transformArrayCloneCall(TR::TreeTop *callTree, OMR::
 
    callTree->insertBefore(TR::TreeTop::create(comp(), TR::Node::create(callNode, TR::treetop, 1, newArray)));
    callTree->insertBefore(TR::TreeTop::create(comp(), TR::Node::create(callNode, TR::treetop, 1, arraycopy)));
+
+   // Flush after the arraycopy so that the cloned array appears identical to the original before it's made
+   // visible to other threads, and because the obj allocation is allowed to use non-zeroed TLH, we need to
+   // make sure no thread sees stale memory contents from the array element section.
+   if (cg()->getEnforceStoreOrder())
+      {
+      TR::Node *allocationFence = TR::Node::createAllocationFence(newArray, newArray);
+      callTree->insertBefore(TR::TreeTop::create(comp(), allocationFence));
+      }
    }
 
 void OMR::ValuePropagation::transformConverterCall(TR::TreeTop *callTree)