-- Jesse Brandeburg > On Feb 13, 2024, at 3:52 PM, Bjorn Helgaas <helgaas@xxxxxxxxxx> wrote: > > Just a heads-up about an ixgbe probe failure seen with Proxmox 8. I > suspect this is a PCI core problem, probably not an ixgbe problem. > > The ixgbe device logs an Advisory Non-Fatal Error and it seems like > subsequent reads from the device return ~0: > > pcieport 0000:00:03.1: AER: Corrected error received: 0000:05:00.0 Why does the user or bios configure corrected errors as fatal or requiring a reset? Seems like a self inflicted wound. > pci 0000:05:00.0: PCIe Bus Error: severity=Corrected, type=Transaction Layer, (Receiver ID) > pci 0000:05:00.0: device [8086:1563] error status/mask=00002000/00000000 > pci 0000:05:00.0: [13] NonFatalErr > > ixgbe 0000:05:00.0: enabling device (0000 -> 0002) > ixgbe 0000:05:00.0: Adapter removed > > The user report is at > https://forum.proxmox.com/threads/proxmox-8-kernel-6-2-16-4-pve-ixgbe-driver-fails-to-load-due-to-pci-device-probing-failure.131203/post-633851. > > I opened a bugzilla with complete dmesg log at > https://bugzilla.kernel.org/show_bug.cgi?id=218491 with some > speculation about what might have caused this, e.g., an ACS > configuration error or something. It's lame, I know, so this is just > a shot in the dark. I’ll look a little more at this tomorrow. I remember lots of inconsistent behaviors around this stuff with early ixgbe and when this PCI capabilities was first enabled. Most of these issues have been resolved since and I haven’t heard of one for a long time. > > Bjorn >
