Move remote dealloc to slow path

It is only ever called in a tail position, so slow path means it is just
a jump, but improves the local deallocation fast path's codegen
considerably.

src/mem/alloc.h

@@ -1370,15 +1370,8 @@ namespace snmalloc
       large_allocator.dealloc(slab, large_class);
     }
 
-#if defined(__GNUC__) && !defined(__clang__) && !defined(__OPTIMIZE__)
+    SNMALLOC_SLOW_PATH
-    // Don't force this to be always inlined in debug builds with GCC, because
+    void remote_dealloc(RemoteAllocator* target, void* p, sizeclass_t sizeclass)
-    // it will fail and then raise an error.
-    inline
-#else
-    SNMALLOC_FAST_PATH
-#endif
-      void
-      remote_dealloc(RemoteAllocator* target, void* p, sizeclass_t sizeclass)
     {
       MEASURE_TIME(remote_dealloc, 4, 16);
       assert(target->id() != id());