Linux Ethtool Network Hardware Diagnostics

The ethtool utility is a standard Linux command-line tool designed to query, control, and diagnose network interface controllers (NICs) and their associated device drivers. It provides systems administrators with direct access to the physical and data link layer settings of a network device, making it an essential diagnostic tool for identifying hardware faults, performance bottlenecks, and link-level negotiation failures.

Inspecting Physical Link Status and Capabilities

At its most fundamental diagnostic level, ethtool reveals whether a physical connection exists and displays the negotiated parameters between the NIC and the switch. By running ethtool [interface_name], you can immediately verify:

• Link Detection: Confirms whether the interface detects an active carrier signal (Link detected: yes/no), isolating physical cabling or transceiver issues.
• Speed and Duplex: Displays the operational speed (e.g., 1000Mb/s, 10000Mb/s) and duplex mode (Full/Half).
• Auto-Negotiation: Shows whether auto-negotiation is enabled and supported by both the local interface and the link partner, helping identify duplex mismatches that cause packet loss.

Analyzing Hardware and Driver Statistics

Hardware degradation or physical transmission errors rarely show up in high-level networking tools like ping or ip. The command ethtool -S [interface_name] queries the NIC’s internal counters directly from the hardware registers. This output helps identify:

• CRC and Frame Errors: High counts usually indicate damaged cabling, faulty transceivers, or electrical interference.
• Dropped Packets and Buffer Overruns: Identifies whether the network card is dropping packets at the hardware FIFO buffer level before the Linux kernel can process them.

Verifying Driver and Firmware Versions

Hardware issues are often tied to outdated firmware or driver bugs. Running ethtool -i [interface_name] displays device driver metadata, including:

• The driver name and version currently loaded by the kernel.
• The firmware version running on the network card itself.
• The PCIe bus address of the device, which helps verify that the card is seated correctly and recognized properly by the system bus.

Executing Hardware Self-Tests

Many enterprise-grade NICs support integrated diagnostic routines. Using ethtool -t [interface_name], administrators can trigger hardware self-tests. These tests evaluate internal registers, memory integrity, and loopback functionality, providing an automated pass/fail report on the physical health of the controller.

Locating Physical Interfaces

In multi-port server environments, identifying which physical port maps to a logical interface name can be difficult. The diagnostic command ethtool -p [interface_name] [seconds] physically blinks the LED on the designated port, allowing technicians to locate the correct port in a server rack for cable swaps or hardware replacements.

Monitoring and Tuning Buffers and Offloads

ethtool diagnoses hardware-level throughput limitations by exposing ring buffer sizes (ethtool -g) and protocol offload features (ethtool -k). If a system drops packets under high loads, examining these settings reveals whether the interface’s receive (RX) and transmit (TX) ring buffers are exhausted, or if hardware checksum offloading, Large Receive Offload (LRO), or TCP Segmentation Offload (TSO) are failing or misconfigured.