SWDEV-257789 - Initial change to skip kernel arg copy

The optimization is controlled with ROCR_SKIP_KERNEL_ARG_COPY.
This is initial check-in for experiments. Extra changes are
necessary for full support:
- handle graph capture with the original sysmem alloc
- avoid memobject references, otherwise there is a race condition with
reusage of the arg buffer
- Remove arg setup from hip

Change-Id: Ib0af710f93e79834711fa4049a7c66093711e68b

rocclr/device/device.hpp

@@ -1214,6 +1214,9 @@ class VirtualDevice : public amd::HeapObject {
   virtual void submitStreamOperation(amd::StreamOperationCommand& cmd) { ShouldNotReachHere(); }
 
   virtual void profilerAttach(bool enable) = 0;
+
+  virtual address allocKernelArguments(size_t size, size_t alignment) { return nullptr; }
+
   //! Get the blit manager object
   device::BlitManager& blitMgr() const { return *blitMgr_; }