feat(fr): Handle FlowControl messages for large messages

Co-authored-by: Tobias Specht <tobias.specht@aisec.fraunhofer.de>

Author: rumpelsepp

src/gallia/transports/flexray_vector.py

@@ -421,6 +421,19 @@ async def read_tp_frame(self) -> FlexRayTPFrame:
         logger.trace("read FlexRayTPFrame %s", repr(frame))
         return frame
 
+    def _require_fc_frame(self, block_size: int, read_bytes: int) -> bool:
+        # 6 bytes already read in first frame.
+        return ((read_bytes - 6) & block_size) == 0
+
+    async def _send_flow_control_frame(self) -> None:
+        block_size = self.config.fc_block_size
+        fc_frame = FlexRayTPFlowControlFrame(
+            flag=FlexRayTPFlowControlFlag.CONTINUE_TO_SEND,
+            separation_time=self.config.fc_separation_time,
+            block_size=block_size,
+        )
+        await self.write_tp_frame(fc_frame)
+
     async def _handle_fragmented(self, expected_len: int) -> bytes:
         # 6 bytes already read in first frame.
         # Headersize is 2 byte.
@@ -431,6 +444,10 @@ async def _handle_fragmented(self, expected_len: int) -> bytes:
         while read_bytes < expected_len:
             # Reordering is not implemented.
             logger.debug(f"expected_len: {expected_len}; read_bytes: {read_bytes}")
+
+            if self._require_fc_frame(self.config.fc_block_size, read_bytes):
+                await self._send_flow_control_frame()
+
             frame = await self.read_tp_frame()
             if not isinstance(frame, FlexRayTPConsecutiveFrame):
                 raise RuntimeError(f"expected consecutive frame, got: {frame}")
@@ -455,15 +472,6 @@ async def read_unsafe(
                 case FlexRayTPSingleFrame():
                     return frame.data
                 case FlexRayTPFirstFrame():
-                    fc_frame = FlexRayTPFlowControlFrame(
-                        flag=FlexRayTPFlowControlFlag.CONTINUE_TO_SEND,
-                        separation_time=self.config.fc_separation_time,
-                        # TODO: send again after block_size number of frames is read.
-                        # Maybe move sending the flow control frame into the
-                        # _handle_fragmented() function and create a loop.
-                        block_size=self.config.fc_block_size,
-                    )
-                    await self.write_tp_frame(fc_frame)
                     data = frame.data + await self._handle_fragmented(frame.size)
                     data = data[: frame.size]
                     logger.debug("read data: %s", data.hex())