#include <stdio.h>
#include <sys/time.h>
#include <string.h>
#include <stdlib.h>

// includes, project
#include <driver_types.h>
#include <cuda_runtime.h>
#include <cuda.h>
#include "helper_cuda.h"
#include "pthread.h"

#include "context.h"

/*
struct Context {
    int next;
    struct Worker* worker;
    struct TaskManager* taskManager;
    int codeNum;
    void  (**code) (struct Context*);
    void* heapStart;
    void* heap;
    long heapLimit;
    int dataNum;
    int idgCount; //number of waiting dataGear
    int idg;
    int maxIdg;
    int odg;
    int maxOdg;
    int workerId;
    struct Context* task;
    struct Queue* tasks;
    int num_exec;
    CUmodule module;
    CUfunction function;
    union Data **data;

    // multi dimension parameter
    int iterate;
    struct Iterator* iterator;
};

struct CUDAWorker {
    CUdevice device;
    CUcontext cuCtx;
    pthread_t thread;
    struct Context* context;
    int id;
    struct Queue* tasks;
    int runFlag;
    int next;
    int numStream;
    CUstream *stream;
} CUDAWorker;

struct LoopCounter {
    int i;
} LoopCounter;

struct Array {
    int size;
    int index;
    int prefix;
    int* array;
} Array;
*/

void cudaInit(struct CUDAWorker *cudaWorker,int phase, int deviceNum) {
    // initialize and load kernel
    cudaWorker->numStream = 1; // number of stream
    //    cudaWorker->stream = NEWN(cudaWorker->numStream, CUstream );
    if (phase==0)
        checkCudaErrors(cuInit(0));
    if (phase==0)
        checkCudaErrors(cuDeviceGet(&cudaWorker->device, deviceNum));
    if (phase==0)
        checkCudaErrors(cuCtxCreate(&cudaWorker->cuCtx, CU_CTX_SCHED_SPIN, cudaWorker->device));
    //    if (cudaWorker->num_stream) {
    //        for (int i=0;i<cudaWorker->num_stream;i++)
    //            checkCudaErrors(cuStreamCreate(&cudaWorker->stream[i],0));
    //    }
    printf("cuda Init: Done\n");
}

void cudaLoadFunction(struct Context* context, char* filename, char* function) {
    checkCudaErrors(cuModuleLoad(&context->module, filename));
    checkCudaErrors(cuModuleGetFunction(&context->function, context->module, function));
}

void cudaShutdown(struct CUDAWorker *worker) {
    //    for (int i=0;i<worker->num_stream;i++)
    //        checkCudaErrors(cuStreamDestroy(worker->stream[i]));
    checkCudaErrors(cuCtxDestroy(worker->cuCtx));
}