initialization of gpu utilization for multiple gpus looks incorrect?

[nishitnshah]

int init_gpu_device_utilization(){
    int i,dev;
    ensure_initialized();
    lock_shrreg();
    for (i=0;i<region_info.shared_region->proc_num;i++){
        for (dev=0;dev<CUDA_DEVICE_MAX_COUNT;dev++){
            region_info.shared_region->procs[i].device_util[dev].sm_util = 0;
            region_info.shared_region->procs[i].monitorused[dev] = 0;
            break;
        }
    }
    unlock_shrreg();
    return 1;
}

The break in the for loop looks buggy? It will only initialize the device 0 and proc 0. What about other procs and devices?