A forced remove and rescan produced the following output in the system log: bb:~ # echo 1 >/sys/devices/pci0000:00/0000:00:01.2/0000:02:00.0/0000:03:04.0/0000:06:00.1/remove Jun 18 17:34:57 bb kernel: pci 0000:06:00.1: Removing from iommu group 27 bb:~ # echo 1 >/sys/devices/pci0000:00/0000:00:01.2/0000:02:00.0/0000:03:04.0/rescan Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: [10ec:8168] type 00 class 0x020000 Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: reg 0x10: [io 0xd800-0xd8ff] Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: reg 0x18: [mem 0xfc514000-0xfc514fff 64bit] Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: reg 0x20: [mem 0xfc508000-0xfc50bfff 64bit] Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: supports D1 D2 Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: PME# supported from D0 D1 D2 D3hot D3cold Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: Adding to iommu group 27 Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: BAR 4: assigned [mem 0xfc508000-0xfc50bfff 64bit] Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: BAR 2: assigned [mem 0xfc514000-0xfc514fff 64bit] Jun 18 17:36:47 bb kernel: pci 0000:06:00.1: BAR 0: assigned [io 0xd800-0xd8ff] Jun 18 17:36:47 bb kernel: r8169 0000:06:00.1: unknown chip XID 54a Conclusion: Rescan does not help, it just reproduces the "unknown chip" error I mentioned earlier. I also verified that I am in fact operating on the correct device: bb:~ # lspci|grep RTL811 06:00.1 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 1a)