Merge branch 4.x

modules/core/src/copy.cpp

@@ -898,14 +898,14 @@ void copyMakeBorder_8u( const uchar* src, size_t srcstep, cv::Size srcroi,
     }
 
     dstroi.width *= elemSize;
-    dst += dststep*top;
 
     for( i = 0; i < top; i++ )
     {
         j = cv::borderInterpolate(i - top, srcroi.height, borderType);
-        memcpy(dst + (i - top)*dststep, dst + j*dststep, dstroi.width);
+        memcpy(dst + i*dststep, dst + (top+j)*dststep, dstroi.width);
     }
 
+    dst += dststep*top;
     for( i = 0; i < bottom; i++ )
     {
         j = cv::borderInterpolate(i + srcroi.height, srcroi.height, borderType);

modules/core/src/hal_internal.cpp

@@ -66,6 +66,7 @@
 
 #if defined(__clang__) && defined(__has_feature)
 #if __has_feature(memory_sanitizer)
+#include <sanitizer/msan_interface.h>
 #define CV_ANNOTATE_MEMORY_IS_INITIALIZED(address, size) \
 __msan_unpoison(address, size)
 #endif

modules/core/src/matrix.cpp

@@ -1158,11 +1158,11 @@ Mat& Mat::adjustROI( int dtop, int dbottom, int dleft, int dright )
         std::swap(col1, col2);
 
     if (dims == 1) {
-        data += (col1 - ofs.x)*esz;
+        data += (col1 - ofs.x)*(std::ptrdiff_t)esz;
         cols = col2 - col1;
         size.p[0] = cols;
     } else {
-        data += (row1 - ofs.y)*step + (col1 - ofs.x)*esz;
+        data += (row1 - ofs.y)*(std::ptrdiff_t)step + (col1 - ofs.x)*(std::ptrdiff_t)esz;
         rows = row2 - row1; cols = col2 - col1;
         size.p[0] = rows; size.p[1] = cols;
         updateContinuityFlag();

modules/core/src/parallel_impl.cpp

@@ -580,8 +580,11 @@ void ThreadPool::run(const Range& range, const ParallelLoopBody& body, double ns
             pthread_mutex_unlock(&mutex);
 
             CV_LOG_VERBOSE(NULL, 5, "MainThread: wake worker threads...");
-            for (size_t i = 0; i < threads.size(); ++i)
+            size_t num_threads_to_wake = std::min(static_cast<size_t>(range.size()), threads.size());
+            for (size_t i = 0; i < num_threads_to_wake; ++i)
             {
+                if (job->current_task >= job->range.size())
+                    break;
                 WorkerThread& thread = *(threads[i].get());
                 if (
 #if defined(__clang__) && defined(__has_feature)