Simulink TCP/IP Receive Data Used 120s (2 Steps) Late in HIL Control (Python Server)

I'm setting up a HIL simulation where a Python script on a Raspberry Pi calculates control commands based on data sent from my Simulink plant model running on a PC . The Simulink model has a fundamental step time of 60 seconds.

Setup & Communication:

Simulink uses TCP/IP Send to send sensor data (20 doubles) to Python.

Python server receives data, performs calculations (~1-2 seconds), and sends back control commands (2 doubles) using client.sendall() over the same connection.

Simulink uses TCP/IP Receive to get these 2 doubles from the Python server

The received Buy/Sell commands are intended to be used by the simulink model in the current simulation step.

I've observed a consistent 120-second (exactly two simulation steps) delay between when Python sends the control commands for time step i and when those commands affect the simulation results in Simulink, which appears to happen during time step i+2.

Python is Fast: Python logs confirm it calculates and executes sendall very quickly (~1-2 seconds) after receiving data for time step i. Logs show "Successfully executed sendall." promptly.

CP/IP Receive Block Configuration (Current Settings):

Data size: [1 2]

Source Data type: double

Byte order: LittleEndian

Enable blocking mode: tried both (enabled and disabled)

Block sample time: -1 or 60

Troubleshooting/Checks:

Confirmed Python sends promptly.

Confirmed basic network connectivity.

My Question:

What could be causing this specific and consistent two-step (120s) delay between Python sending data and Simulink using it?

Given the TCP/IP Receive settings, are there other common configuration issues that manifest as a multi-step delay for TCP/IP received data?

I need the commands calculated for time step i to be used during the simulation of Hour i. How can I configure the blocks or diagnose the model further to eliminate this 120s delay?

Thanks for any insights!