Fix for kernel synchronization

The completion future of a particular kernel is lost if there are
multiple kernels in the stream. This can cause a racing condition where
the signal associated with the unreferenced completion_future might get
released by hcc runtime.

include/hcc_detail/hip_hcc.h

@@ -440,9 +440,13 @@ public:
     hc::accelerator_view        _av;
     unsigned                    _flags;
 
-private: // Critical Data.  THis MUST be accessed through LockedAccessor_StreamCrit_t
+private:
+    // Critical Data.  THis MUST be accessed through LockedAccessor_StreamCrit_t
     ihipStreamCritical_t        _criticalData;
 
+    // Array of dependency completion_future.
+    std::vector<hc::completion_future> _depFutures;
+
 private:
     void                        enqueueBarrier(hsa_queue_t* queue, ihipSignal_t *depSignal);
     void                        waitCopy(LockedAccessor_StreamCrit_t &crit, ihipSignal_t *signal);