Transfer Linked list from CPU to GPU

I want to transfer a Quadtree, which is a linked list built in CPU using C program to the GPU. I am using CUDA 3.0. I read some posts on the internet that transferring linked list to GPU is quite hard and instead, I could flatten the list to an array and transfer the array to GPU.

But is there a way I could make my quadtree common to CPU and GPU and still use pointers on the GPU side to access the quad tree nodes?

Or if that is not possible , what else can I do to transfer the linked list to GPU? Can anyone point me to a detailed document on this?

Thanks