OCL and custom OpenCl kernel

edit

Basically there seems to be no way to use custom kernels with OCL except the openCLExecuteKernelInterop() function which has no documentation and is tagged as "Deprecated"!

I also tried the code that abarral suggested in his question but i couldn't make it run due an unknown error in openCLExecuteKernelInterop(according to gdb).

So is there a way to avoid writing 10K lines of host code for a simple kernel? In the case that answer is no, is it possible to use oclMat with OpenCL and avoid manually handling memory buffers and just use .download() & .upload() methods?

Edit: This is the exact code that abarral has used in his question:

#include <iostream>
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <opencv2/ocl/ocl.hpp>


int main(int _argc, const char** _argv)
{
    cv::ocl::DevicesInfo devInfo;
    int res = cv::ocl::getOpenCLDevices(devInfo);
    if(res == 0)
    {
        std::cout << "There is no OPENCL Here !" << std::endl;
    }else
    {
        for(int i = 0 ; i < devInfo.size() ; ++i)
        {
            std::cout << "Device : " << devInfo[i]->deviceName 
            << " is present" << std::endl;
        }
    }
    cv::ocl::setDevice(devInfo[0]);        // select device to use
    std::cout << CV_VERSION_EPOCH << "." 
    << CV_VERSION_MAJOR << "." << CV_VERSION_MINOR << std::endl;
    const char *KernelSource = "\n" \
    "__kernel void square(                       \n" \
    "   __global uchar* input,                   \n" \
    "   __global uchar* output)                  \n" \
    "{                                           \n" \
    "   int i = get_global_id(0);                \n" \
    "   output[i] = input[i] * input[i];         \n" \
    "}\n";
    cv::ocl::ProgramSource src("square", KernelSource);
    std::size_t globalThreads[3]={1,0,0};
    std::size_t localThreads[3]={5,5,0};
    cv::ocl::oclMat source(cv::Size(500,500), CV_8UC1);
    cv::ocl::oclMat dest(cv::Size(500,500), CV_8UC1);

    std::vector<std::pair<size_t , const void *> > args;
    args.push_back( std::make_pair( 4, (void *) &source.data ));
    args.push_back( std::make_pair( 4, (void *) &dest.data ));

    cv::ocl::openCLExecuteKernelInterop(cv::ocl::Context::getContext(), src, "square",
                                     globalThreads, localThreads, args, -1, -1, "");
    return 0;
}

It gets terminated by "receiving signal 8 (Floating point exception)". GDB brings:

Program received signal SIGFPE, Arithmetic exception. 0x00007ffff469cc6b in cv::ocl::openCLExecuteKernelInterop(cv::ocl::Context, cv::ocl::ProgramSource const&, std::string, unsigned long, unsigned long, std::vector<std::pair<unsigned long,="" void="" const*="">, std::allocator<std::pair<unsigned long,="" void="" const*=""> > >&, int, int, char const) () from /usr/lib/libopencv_ocl.so.2.4