如何使用CUDA执行struct的深度复制？

#include<stdio.h>
#include <stdlib.h>
#include <cuda_runtime.h>

typedef struct {
    int x;
    int y;
} Point;

__global__ void copy_point(Point *dst, Point *src) {
    *dst = *src;
}

int main() {
    Point *h_src, *h_dst;
    Point *d_src, *d_dst;

    // 分配主机内存
    h_src = (Point *)malloc(sizeof(Point));
    h_dst = (Point *)malloc(sizeof(Point));

    // 初始化主机内存
    h_src->x = 1;
    h_src->y = 2;

    // 分配设备内存
    cudaMalloc((void **)&d_src, sizeof(Point));
    cudaMalloc((void **)&d_dst, sizeof(Point));

    // 将主机内存复制到设备内存
    cudaMemcpy(d_src, h_src, sizeof(Point), cudaMemcpyHostToDevice);

    // 调用核函数执行深度复制
    copy_point<<<1, 1>>>(d_dst, d_src);

    // 将设备内存复制回主机内存
    cudaMemcpy(h_dst, d_dst, sizeof(Point), cudaMemcpyDeviceToHost);

    // 检查结果
    printf("src: (%d, %d)\n", h_src->x, h_src->y);
    printf("dst: (%d, %d)\n", h_dst->x, h_dst->y);

    // 释放内存
    free(h_src);
    free(h_dst);
    cudaFree(d_src);
    cudaFree(d_dst);

    return 0;
}