diff --git a/ggml/src/ggml-qnn/qnn-lib.hpp b/ggml/src/ggml-qnn/qnn-lib.hpp
index 7307c9f63..a676f9895 100644
--- a/ggml/src/ggml-qnn/qnn-lib.hpp
+++ b/ggml/src/ggml-qnn/qnn-lib.hpp
@@ -36,6 +36,11 @@ inline int dl_unload(dl_handler_t handle) { return dlclose(handle); }
 
 inline const char *dl_error() { return dlerror(); }
 
+template <typename Fn>
+Fn dl_sym_typed(dl_handler_t handle, const std::string &function_name) {
+    return reinterpret_cast<Fn>(dl_sym(handle, function_name));
+}
+
 // =================================================================================================
 //
 // wrapper class of Qualcomm QNN(Qualcomm Neural Network, aka Qualcomm AI Engine Direct) SDK
@@ -722,8 +727,8 @@ private:
             return 1;
         }
 
-        auto *get_providers = reinterpret_cast<qnn::pfn_qnnsysteminterface_getproviders *>(
-            dl_sym(system_lib_handle, "QnnSystemInterface_getProviders"));
+        auto *get_providers = dl_sym_typed<qnn::pfn_qnnsysteminterface_getproviders *>(
+            system_lib_handle, "QnnSystemInterface_getProviders");
         if (!get_providers) {
             QNN_LOG_WARN("can not load QNN symbol QnnSystemInterface_getProviders: %s\n", dl_error());
             return 2;
@@ -784,8 +789,8 @@ private:
             return 1;
         }
 
-        auto get_providers = qnn::load_qnn_functionpointers<qnn::pfn_qnninterface_getproviders *>(
-            lib_handle, "QnnInterface_getProviders");
+        auto get_providers =
+            qnn::dl_sym_typed<qnn::pfn_qnninterface_getproviders *>(lib_handle, "QnnInterface_getProviders");
         if (!get_providers) {
             QNN_LOG_WARN("can not load symbol QnnInterface_getProviders : %s", dl_error());
             return 2;
diff --git a/ggml/src/ggml-qnn/utils.hpp b/ggml/src/ggml-qnn/utils.hpp
index 4a01347d0..66c3eeba4 100644
--- a/ggml/src/ggml-qnn/utils.hpp
+++ b/ggml/src/ggml-qnn/utils.hpp
@@ -24,11 +24,6 @@ uint32_t get_ggml_tensor_data_size(const ggml_tensor *tensor);
 
 const char *opname_from_ggmlop(enum ggml_op ggmlop);
 
-template <typename Fn>
-Fn load_qnn_functionpointers(void *handle, const char *function_name) {
-    return reinterpret_cast<Fn>(dlsym(handle, function_name));
-}
-
 inline int validate_tensor_version(const Qnn_Tensor_t &tensor) {
     if (tensor.version != QNN_TENSOR_VERSION_1) {
         QNN_LOG_WARN("validate_tensor_version() tensor %s, got unsupported version %d\n", tensor.v1.name,