Linux e1000e base driver for Intel Gigabit Ethernet Network Connection

Users with a problem Ethernet connection are not unheard of, but I'd start with changes to Cables and Switches before pointing the finger at the kernel driver. LE uses the upstream in-kernel driver - which is also used by 10's of millions of other Intel devices. It's not impossible for the driver to have an issue, but if you do want to point the finger in that direction you will need to convince us by sharing some proper technical evidence of it being the source of problems! NB: In all cases we will not update to the not-in-kernel vendor driver, it would be a bad move.

Quote from vhgroup

I have finally narrowed down my Kodi NAS Playback issues to a problematic Ethernet connection

could you describe your problem a bit more ?

very likely its not the driver

Quote from vhgroup

Anyway, may I ask, what is the harm in shipping an improved driver with multiple bug fixes?

Test an LE11 nightly from https://test.libreelec.net/ and you will be using the latest LTS kernel driver that Intel has contributed fixes to. If we switched from the in-kernel driver frequently maintained by Intel to some forgotten out-of-tree kernel driver that Intel has not touched in two years; you can expect more breakage than fixage.

Quote from CvH

very likely its not the driver

+1

running e1000e over years with different kernels without errors, esp. connection problems.

same goes for my NUC (see signature)

surfing the kernel bugzilla (e1000e): bugs are all elder and more PCI-e (?) power save related AND should be fixed since, otherwise there should be newer reports.

couldn't figure what's the last version of the e1000e driver in an kernel tree is.

any hints ?

if it related to the date in the copyright string in the driver files: the newest date is "2018" (Kernel 5.16)

I guess a LE log file could show some more what's up with connection problem the OP is seeing

vhgroup - can you please share a full dmesg?

- Should show the model of NUC.

- What are the errors shown in the logs?

What is the setting of AMT in the BIOS? Reportedly the below can change things.

BIOS setttings -> System Management -> Intel AMT Capability
switch it from "Restrict MEBx Access" to "Disabled"

To enable debug logs on your e1000e card (not suggesting this as a first pass:)

echo "module e1000e +flmpt" > /sys/kernel/debug/dynamic_debug/control

echo 8 > /proc/sys/kernel/printk

Then check journalctl -f and see what happens when you are having the failure. (It is noisy)

Quote from vhgroup

Yet the version number returned by ethtool is 5.10.76? So now I am confused.

that is the kernel version not the driver version

re-check with "uname -r"

Bump to the LE11 nightly, because if anything is fixed that's great, and if not fixed it's a current kernel to investigate from wheras the 5.10.x kernel used in LE10 releases is recent but also ageing .. Linux 5.16 was released last Sunday.

Quote from vhgroup

Version 3.8.4 seems to be the latest version, released almost 2 years ago!

according to this document (date 2020):

Release Notes (lan-25.6-release_notes.pdf

the linux driver version for the 200 LAN Series is 5.3.x

Quote from chewitt

Linux 5.16 was released last Sunday.

... and it seems it might hit nightly tomorrow ...

Quote from vhgroup

I believe the following link is very relevant: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1766377

If you believe it's relevant have you tried the workarounds it contains? e.g. "ethtool -K eth0 gso off gro off tso off" .. If you shared logs (as asked multiple times) we'd be able to confirm your box produces the same hang messages. No logs = No problem.

NB: LE11 nightlies are here: https://test.libreelec.net/ .. If you want to test safely, stop Kodi, rename /storage/.kodi to .kodi-old and then drop the udpate .tar file in /storage/.update and reboot. You will start with a clean Kodi install - which means we don't "upgrade" all your existing config to LE11 and the downgrade is simply the reverse process; stop Kodi, move folders around so the old folder so the active one and drop the older 'update' file in /storage/.update and reboot again. It's messing .. but low drama.

Hi vhgroup - I see you have a NUC6 (not sure on the model) - though has a different Ethernet Chip - but we have the same driver. Mine doesn’t experience th3 same issue that you have described.

Dumb question…. Did you have wireless enabled at the same time as Ethernet?

nuc6:~ # dmesg | grep -e DMI: -e e1000

[ 0.000000] DMI: /NUC6i5SYB, BIOS SYSKLi35.86A.0073.2020.0909.1625 09/09/2020

[ 1.263095] e1000: Intel(R) PRO/1000 Network Driver

[ 1.263096] e1000: Copyright (c) 1999-2006 Intel Corporation.

[ 1.263108] e1000e: Intel(R) PRO/1000 Network Driver

[ 1.263109] e1000e: Copyright(c) 1999 - 2015 Intel Corporation.

[ 1.263306] e1000e 0000:00:1f.6: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode

[ 1.464787] e1000e 0000:00:1f.6 0000:00:1f.6 (uninitialized): registered PHC clock

[ 1.532147] e1000e 0000:00:1f.6 eth0: (PCI Express:2.5GT/s:Width x1)

[ 1.532153] e1000e 0000:00:1f.6 eth0: Intel(R) PRO/1000 Network Connection

[ 1.532223] e1000e 0000:00:1f.6 eth0: MAC: 12, PHY: 12, PBA No: FFFFFF-0FF

[ 26.570861] e1000e 0000:00:1f.6 eth0: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None

nuc6:~ # lspci -v -s 00:1f.6

00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection I219-V (rev 21)

DeviceName: LAN

Subsystem: Intel Corporation Ethernet Connection I219-V

Flags: bus master, fast devsel, latency 0, IRQ 125

Memory at df100000 (32-bit, non-prefetchable) [size=128K]

Capabilities: [c8] Power Management version 3

Capabilities: [d0] MSI: Enable+ Count=1/1 Maskable- 64bit+

Capabilities: [e0] PCI Advanced Features

Kernel driver in use: e1000e