Add missing stride in cuda::LinearFilter::apply - #3163

[cudawarped]

This aplies the fix identified by @yl070360 in #3163.

The stride information is currently missing inside cuda::LinearFilter::apply() producing incorrect results when GpuMat's are constructed using pointers to pre-allocated device memory.

The current tests should to be sufficient to check this because they are wrongly passing at the moment due to another bug (see below) in the construction of GpuMat's which can be fixed once this is merged.

GpuMat Bug
GpuMat::dataend wrongly points to the last element of an alocated row instead of the last useable element of the ROI unless the GpuMat is constructed from a pointer to pre-allocated device memory.
This means that locateROI(wholeSize,ofs) currently returns the stride in wholeSize instead of the width of the ROI.

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

• I agree to contribute to the project under Apache 2 License.
• To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
• The PR is proposed to the proper branch
• There is a reference to the original bug report and related work
• There is accuracy test, performance test and test data in opencv_extra repository, if applicable
  Patch to opencv_extra has the same branch name.
• The feature is well documented and sample code can be built with the project CMake

force_builders=Custom
buildworker:Custom=linux-4,linux-6
build_image:Custom=ubuntu-cuda:18.04

[alalek]

Merge branch 'master' of ...

Could you please rebase on upstream/4.x to eliminate unrelated commits? (keep 1 commit in PR)

(where upstream is configured by following this GitHub guide and fetched (git fetch upstream))