Ethernet interface understanding in Drive AGX thor HW

DRIVE OS Version: Provide DRIVE OS version. Example: 7.2.5

I would like to connect the 100B T1 Link of thor HW to Another controller

How can i identify which port i can connect and how to set the ethernet configuration .

Thanks ,

What is difference between MGBE interfaces and the ETHER_QOS controllers

Disclaimer: Synthesized with AI assistance; technical details should be verified against official documentation.

Hi @nandhini.ravikumar1,

Here is a breakdown of how Ethernet is routed on the DRIVE AGX Thor platform to help you connect your 100BASE-T1 link:

1. Difference Between MGBE and EQOS (ETHER_QOS)

Both of these are hardware Ethernet MACs integrated directly into the Tegra SoC, but they serve different roles:

  • MGBE (Multi-Gigabit Ethernet): These are the high-speed MACs (typically supporting up to 10G/25G). On the DRIVE AGX platform, the MGBE interfaces are usually connected internally to the on-board automotive Ethernet switch (e.g., Marvell) as high-bandwidth backhaul links, or routed to external high-speed connectors (like QSFP).

  • EQOS (Ethernet Quality of Service): This is typically a 1-Gigabit (1GbE) MAC. It is usually exposed as eqos_0 in Linux and is the primary interface used for platform management, flashing, and connecting to a standard RJ45 network. It features hardware-level QoS, which is why it is often used for PTP (Precision Time Protocol) syncing.

2. Connecting a 100BASE-T1 Link

On DRIVE AGX boards, Automotive Ethernet interfaces (like 100BASE-T1 and 1000BASE-T1) do not connect directly to the Tegra SoC’s EQOS or MGBE pins. Instead, they connect to the on-board Automotive Ethernet Switch.

To connect and configure your external controller:

  1. Identify the Physical Port: You will need to check the exact J- or P- connector pins assigned to 100BASE-T1 on your specific carrier board topology.

  2. Switch Configuration: Because the port sits on the managed switch, you cannot simply configure an IP address on eqos_0. The switch firmware must be configured to route traffic from that specific 100BASE-T1 PHY port to one of the Tegra’s internal interfaces (usually via an MGBE link using VLAN tagging).

  3. Linux Setup: Once the switch routing is confirmed, the port will appear in Linux as a virtual interface or a VLAN interface mapped to the backhaul MAC. You can then assign it a static IP using standard netplan or ifconfig commands.

Official Documentation Reference

Because DRIVE OS 7.2.5 requires developer portal access, you can find the exact port mappings by logging into your NVIDIA Developer account and navigating to the “Network Topology” section of the NVIDIA DriveOS 7.2.5 Linux SDK Early Access Developer Guide:Welcome to the NVIDIA DriveOS 7.2.5.0 Developer Guide — NVIDIA DriveOS 7.2.5 Linux SDK Early Access Developer Guide

That section contains the exact block diagrams and connector pinouts for your hardware.

Thank you @Andrey1984 for your suggestions.

Also, please check k P3960 Network Topology — NVIDIA DriveOS 7.2.5 Linux SDK Early Access Developer Guide table to know 100Mbps port. You need to use 1G dongle to connect to LAN for 100Mbps ports.

@Andrey1984 :Thanks a lot , it is really helpful

@SivaRamaKrishnaNV : Thanks a lot . I will check and will post if i stuck