turn on some instantiates that required for other primitives

modules/gpu/src/cuda/copy_make_border.cu

@@ -105,9 +105,7 @@ namespace cv { namespace gpu { namespace device
         }
 
         template void copyMakeBorder_gpu<uchar, 1>(const PtrStepSzb& src, const PtrStepSzb& dst, int top, int left, int borderMode, const uchar* borderValue, cudaStream_t stream);
-#ifndef OPENCV_TINY_GPU_MODULE
         template void copyMakeBorder_gpu<uchar, 2>(const PtrStepSzb& src, const PtrStepSzb& dst, int top, int left, int borderMode, const uchar* borderValue, cudaStream_t stream);
-#endif
         template void copyMakeBorder_gpu<uchar, 3>(const PtrStepSzb& src, const PtrStepSzb& dst, int top, int left, int borderMode, const uchar* borderValue, cudaStream_t stream);
         template void copyMakeBorder_gpu<uchar, 4>(const PtrStepSzb& src, const PtrStepSzb& dst, int top, int left, int borderMode, const uchar* borderValue, cudaStream_t stream);
 

modules/gpu/src/cuda/element_operations.cu

@@ -2193,15 +2193,17 @@ namespace arithm
         transform((PtrStepSz<T>) src1, (PtrStepSz<T>) dst, cv::gpu::device::bind2nd(minimum<T>(), src2), WithOutMask(), stream);
     }
 
+#ifdef OPENCV_TINY_GPU_MODULE
+    template void minScalar<uchar >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
+    template void minScalar<int   >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
+    template void minScalar<float >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
+#else
     template void minScalar<uchar >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
-#ifndef OPENCV_TINY_GPU_MODULE
     template void minScalar<schar >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
     template void minScalar<ushort>(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
     template void minScalar<short >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
     template void minScalar<int   >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
-#endif
     template void minScalar<float >(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
-#ifndef OPENCV_TINY_GPU_MODULE
     template void minScalar<double>(PtrStepSzb src1, double src2, PtrStepSzb dst, cudaStream_t stream);
 #endif
 }

modules/gpu/src/cuda/matrix_reductions.cu

@@ -975,8 +975,8 @@ namespace minMaxLoc
     template void run<signed char >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
     template void run<unsigned short>(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
     template void run<short >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
-    template void run<int   >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
 #endif
+    template void run<int   >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
     template void run<float >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);
 #ifndef OPENCV_TINY_GPU_MODULE
     template void run<double>(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, int* minloc, int* maxloc, PtrStepb valbuf, PtrStep<unsigned int> locbuf);

modules/gpu/src/element_operations.cpp

@@ -3717,7 +3717,7 @@ void cv::gpu::min(const GpuMat& src, double val, GpuMat& dst, Stream& stream)
         0/*minScalar<signed char>*/,
         0/*minScalar<unsigned short>*/,
         0/*minScalar<short>*/,
-        0/*minScalar<int>*/,
+        minScalar<int>,
         minScalar<float>,
         0/*minScalar<double>*/,
     };
@@ -3728,7 +3728,7 @@ void cv::gpu::min(const GpuMat& src, double val, GpuMat& dst, Stream& stream)
         0/*castScalar<signed char>*/,
         0/*castScalar<unsigned short>*/,
         0/*castScalar<short>*/,
-        0/*castScalar<int>*/,
+        castScalar<int>,
         castScalar<float>,
         0/*castScalar<double>*/,
     };

modules/gpu/src/imgproc.cpp

@@ -339,12 +339,12 @@ void cv::gpu::copyMakeBorder(const GpuMat& src, GpuMat& dst, int top, int bottom
 #ifdef OPENCV_TINY_GPU_MODULE
         static const caller_t callers[6][4] =
         {
-            {   copyMakeBorder_caller<uchar, 1>  ,  0/*copyMakeBorder_caller<uchar, 2>*/,    copyMakeBorder_caller<uchar, 3>  ,    copyMakeBorder_caller<uchar, 4>},
+            {   copyMakeBorder_caller<uchar, 1>  ,  copyMakeBorder_caller<uchar, 2>     ,    copyMakeBorder_caller<uchar, 3>  ,    copyMakeBorder_caller<uchar, 4>},
             {0, 0, 0, 0},
             {0, 0, 0, 0},
             {0, 0, 0, 0},
             {0, 0, 0, 0},
-            {   copyMakeBorder_caller<float, 1>  , 0/*copyMakeBorder_caller<float, 2>*/ ,    copyMakeBorder_caller<float, 3>  ,    copyMakeBorder_caller<float ,4>}
+            {   copyMakeBorder_caller<float, 1>  ,  0/*copyMakeBorder_caller<float, 2>*/,    copyMakeBorder_caller<float, 3>  ,    copyMakeBorder_caller<float ,4>}
         };
 #else
         static const caller_t callers[6][4] =

modules/gpu/src/matrix_reductions.cpp

@@ -523,7 +523,7 @@ void cv::gpu::minMaxLoc(const GpuMat& src, double* minVal, double* maxVal, Point
         0/*::minMaxLoc::run<schar>*/,
         0/*::minMaxLoc::run<ushort>*/,
         0/*::minMaxLoc::run<short>*/,
-        0/*::minMaxLoc::run<int>*/,
+        ::minMaxLoc::run<int>,
         ::minMaxLoc::run<float>,
         0/*::minMaxLoc::run<double>*/,
     };