[C-Api] fix memory leak bug

Since developers can reduce the number of Tensors by calling
ml_tensors_info_set_count(), it causes the memory leak when invoking
ml_tensors_info_free() and ml_tensors_data_destroy(). This patch fixes
that bug by checking all possible elements (i.e. ML_TENSOR_SIZE_LIMIT).

Signed-off-by: Sangjung Woo <sangjung.woo@samsung.com>

diff --git a/api/capi/src/nnstreamer-capi-util.c b/api/capi/src/nnstreamer-capi-util.c
index 432b176..b9b47c5 100644 (file)
--- a/api/capi/src/nnstreamer-capi-util.c
+++ b/api/capi/src/nnstreamer-capi-util.c
@@ -397,7 +397,7 @@ ml_tensors_info_free (ml_tensors_info_s * info)
   if (!info)
     return;
 
-  for (i = 0; i < info->num_tensors; i++) {
+  for (i = 0; i < ML_TENSOR_SIZE_LIMIT; i++) {
     if (info->info[i].name) {
       g_free (info->info[i].name);
       info->info[i].name = NULL;
@@ -421,7 +421,7 @@ ml_tensors_data_destroy (ml_tensors_data_h data)
 
   _data = (ml_tensors_data_s *) data;
 
-  for (i = 0; i < _data->num_tensors; i++) {
+  for (i = 0; i < ML_TENSOR_SIZE_LIMIT; i++) {
     if (_data->tensors[i].tensor) {
       g_free (_data->tensors[i].tensor);
       _data->tensors[i].tensor = NULL;