Map/unmap enqueue fixes [2/n]: CPU operations on limited range

- Curently each non-zerocopy CPU operation on map/unmap make a full copy
  using hostPtr
- This commit adds functionality to select specific range of copy
- Multiple mapping with different size is not supported yet,
  so copy will be made on full range for now. This is for future usage.

Change-Id: I7652e85482ba6fffb2474169447baf9b080dcd1e

runtime/mem_obj/buffer.h

@@ -100,7 +100,8 @@ class Buffer : public MemObj {
 
     static size_t calculateHostPtrSize(const size_t *origin, const size_t *region, size_t rowPitch, size_t slicePitch);
 
-    void *transferDataToHostPtr() override;
+    void transferDataToHostPtr(std::array<size_t, 3> copySize, std::array<size_t, 3> copyOffset) override;
+    void transferDataFromHostPtr(std::array<size_t, 3> copySize, std::array<size_t, 3> copyOffset) override;
 
     bool isReadWriteOnCpuAllowed(cl_bool blocking, cl_uint numEventsInWaitList, void *ptr, size_t size);
 
@@ -125,6 +126,8 @@ class Buffer : public MemObj {
                             bool &allocateMemory,
                             bool &copyMemoryFromHostPtr,
                             MemoryManager *memMngr);
+
+    void transferData(void *dst, void *src, size_t copySize, size_t copyOffset);
 };
 
 template <typename GfxFamily>