How to get the device pointer of a OpenGL texture?

Accelerated Computing CUDA CUDA Programming and Performance
1

My program use NPP library’s “nppiResize_XXX()” function to do high quality image resize,the source & target are OpenGL textures with RGBA internal format,but we know the pointer arguments of function “nppiResize_XXX()” are device pointers,so how can I get the device pointer of a OpenGL texture?

I had use “cudaGraphicsGLRegisterImage()”,“cudaGraphicsMapResources()” & “cudaGraphicsResourceGetMappedMipmappedArray()” runtime apis,but I only get a “cudaMipmappedArray_t” pointer,I confused,what’s the meaning of “cudaMipmappedArray_t” pointer?how can I get the device pointer?