Comment 83 for bug 1521173

I can confirm similar issues on AMD Threadripper.

The issues went away on Ubuntu 17.10 by adding "pci_aspm=off" to grub, but are re-introduced by upgrading to 18.04. Strangely, the error now occurs in both 4.13 and 4.15 kernels.

One solution I found is to set PCIe to 2.0 instead of the default 3.0 in BIOS (but this comes at a slight performance cost).

Do we know if this will be assigned / solved soon?

May 4 13:09:38 TR-Ubuntu kernel: [ 76.552730] pcieport 0000:00:01.1: [12] Replay Timer Timeout
May 4 13:09:38 TR-Ubuntu kernel: [ 76.563746] dpc 0000:00:01.1:pcie010: DPC containment event, status:0x1f00 source:0x0000
May 4 13:09:38 TR-Ubuntu kernel: [ 76.563759] pcieport 0000:00:01.1: AER: Multiple Corrected error received: id=0000
May 4 13:09:38 TR-Ubuntu kernel: [ 76.563788] pcieport 0000:00:01.1: PCIe Bus Error: severity=Corrected, type=Data Link Layer, id=0009(Transmitter ID)
May 4 13:09:38 TR-Ubuntu kernel: [ 76.563790] pcieport 0000:00:01.1: device [1022:1453] error status/mask=00001080/00006000
May 4 13:09:38 TR-Ubuntu kernel: [ 76.563792] pcieport 0000:00:01.1: [ 7] Bad DLLP