GPUTracking: declare the resolve kernel's shared memory as shared (#15860)

The definitions in GPUTPCGMMergerGPU.cxx qualify smem with GPUsharedref(), and
so does every other declaration in the header; this one did not. It makes no
difference where GPUsharedref() is empty, but on Metal the declaration then
takes a generic reference and the definition a threadgroup one, which are
different types.

Author: ktf

GPU/GPUTracking/Merger/GPUTPCGMMergerGPU.h

@@ -81,7 +81,7 @@ class GPUTPCGMMergerResolve : public GPUTPCGMMergerGeneral
   };
 
   template <int32_t iKernel = defaultKernel, typename... Args>
-  GPUd() static void Thread(int32_t nBlocks, int32_t nThreads, int32_t iBlock, int32_t iThread, GPUSharedMemory& smem, processorType& clusterer, Args... args);
+  GPUd() static void Thread(int32_t nBlocks, int32_t nThreads, int32_t iBlock, int32_t iThread, GPUsharedref() GPUSharedMemory& smem, processorType& clusterer, Args... args);
 };
 
 class GPUTPCGMMergerClearLinks : public GPUTPCGMMergerGeneral