Add some notes in the gpu_info implementation

edgeless_node/src/gpu_info/jetson.rs

@@ -189,7 +189,9 @@ pub fn jetson_get_mem_size_gpu() -> i32 {
 /// # Returns
 /// * `i32` - The number of available GPUs or a negative number if the operation fails
 pub fn jetson_get_num_gpus() -> i32 {
-    1 // TODO: implement this
+    // Jetson boards only contain a single GPU
+    // see: https://forums.developer.nvidia.com/t/detect-number-of-gpus/314622
+    1
 }
 
 /// Determines if the system is a Jetson board by checking `/sys/firmware/devicetree/base/model`.

edgeless_node/src/gpu_info/mod.rs

@@ -3,6 +3,17 @@
 
 pub mod jetson;
 
+// =====================================================================================
+// *** IMPORTANT ***
+// To support an additional GPU in the future, follow these steps:
+//
+//  1. Create a new .rs file in the same directory for the newly supported GPU.
+//  2. Update `enum BoardType` to accommodate the new hardware.
+//  3. Implement accessors (getters) for the newly supported hardware in its .rs file,
+//         e.g., is_nvidia_board(), nvidia::nvidia_get_gpu_temp(), etc.
+//  4. Update this source code's API calls to handle the new hardware if it is detected.
+// =====================================================================================
+
 /// Enum to represent different board types.
 /// If you need to support more board types, add them here.
 #[derive(Debug, PartialEq)]