Comment 11 for bug 1758545

I can also report the same behavior, on a Clevo laptop running Ubuntu 20.04 (development version), kernel 5.3.0-24. Clevo does not currently provide BIOS updates for this model, and it is unlikely that they ever will.

The error is as reported by the OP, and repeats several times per second on the standard error, which makes it impossible to use virtual consoles and terminals:

  [timestamp] pcieport 0000:00:1d.4: AER: PCIe Bus Error: severity=Uncorrected (Non-Fatal), type=Transaction Layer, (Requester ID)
  [timestamp] pcieport 0000:00:1d.4: AER: device [8086:9db4] error status/mask=00100000/00010000
  [timestamp] pcieport 0000:00:1d.4: AER: [20] UnsupReq (First)
  [timestamp] pcieport 0000:00:1d.4: AER: TLP Header: 34000000 04000010 00000000 80008000

I have followed the explanation given here for a very similar PCI Express bus error:

  https://unix.stackexchange.com/a/369090

to give the option pci=nommconf to the kernel, in order to disable memory-mapping the configuration space of the PCI device, and instead use the original method of accessing this space through I/O ports, if I understand correctly. This option can be given at startup (press E in GRUB) and then in /etc/default/grub for subsequent boots (GRUB_CMDLINE_LINUX_DEFAULT="quiet splash pci=nommconf").

The option works for me in working around the issue, and can hopefully work for other users whose hardware won't be supported with a BIOS update.

From what I understand, the problem could come from the particular PCI Express device (an Intel PCIe NVMe M.2 SSD, 760p Series, according to lshw), from the PCI Express root controller, or from the interaction between them.