On 3/3/2025 5:30 PM, Paul Menzel wrote:
Dear Linux folks,
On a Dell OptiPlex 7000/0F1HC1, BIOS 1.8.2 12/08/2022 with
00:1f.6 Ethernet controller [0200]: Intel Corporation Ethernet
Connection (17) I219-LM [8086:1a1c] (rev 11)
and Dell OptiPlex 7071/097YXY, BIOS 1.1.2 08/29/2019 with
00:1f.6 Ethernet controller [0200]: Intel Corporation Ethernet
Connection (7) I219-LM [8086:15bb] (rev 10)
we recently observed that randomly auto-negotiation would only result in
a speed of 100 Mb/s or once even 10 Mb/s. (Wake-on-LAN is enabled for
the device, although it does not work for the Dell OptiPlex 7000.)
Today, the Dell OptiPlex 7071 booted and only came up with 10 Mb/s. Only
after re-plugging the cable in both plugs (device and in the ground),
and applying contact spray, it negotiated a speed of 1000 Gb/s. Then
rebooting, it negotiated a speed 1000 Gb/s, but than a minute later, the
link went down, and, 22 seconds later, it came back with 100 Mb/s.
Mar 03 15:35:35 sighup.molgen.mpg.de kernel: Linux version
6.12.11.mx64.479 ([email protected]) (gcc (GCC) 12.3.0, GNU ld
(GNU Binutils) 2.41) #1 SMP PREEMPT_DYNAMIC Fri Jan 24 13:30:47 CET 2025
[…]
Mar 03 15:35:40 sighup.molgen.mpg.de kernel: e1000e 0000:00:1f.6
net00: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
[…]
Mar 03 15:37:31 sighup.molgen.mpg.de kernel: e1000e 0000:00:1f.6
net00: NIC Link is Down
[…]
Mar 03 15:37:53 sighup.molgen.mpg.de kernel: e1000e 0000:00:1f.6
net00: NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx
Then I enabled dynamic debug messages
echo "module e1000e +flmpt" | sudo tee
/sys/kernel/debug/dynamic_debug/control
and tried to force negotiating a speed of 1 Gb/s:
sudo ethtool -s net00 autoneg on speed 1000 duplex full
The Linux kernel messages are attached:
2025-03-03T15:39:00+01:00 sighup sudo: pmenzel : TTY=pts/0 ;
PWD=/home/pmenzel ; USER=root ; COMMAND=/usr/bin/tee
/sys/kernel/debug/dynamic_debug/control
[…]
2025-03-03T15:39:08+01:00 sighup sudo: pmenzel : TTY=pts/0 ;
PWD=/home/pmenzel ; USER=root ; COMMAND=/usr/sbin/ethtool -s net00
autoneg on speed 1000 duplex full
[…]
2025-03-03T15:39:08.768806+01:00 sighup kernel: [ 217.278630]
[1168] e1000e:e1000_reset_hw_ich8lan:4778: e1000e 0000:00:1f.6 net00:
Masking off all interrupts
[…]
2025-03-03T15:39:08.834765+01:00 sighup kernel: [ 217.344955]
[1168] e1000e:e1000e_setup_copper_link:1216: e1000e 0000:00:1f.6 net00:
Unable to establish link!!!
[…]
2025-03-03T15:39:08.840771+01:00 sighup kernel: [ 217.350769]
[1168] e1000e:__e1000_write_phy_reg_hv:2963: e1000e 0000:00:1f.6 net00:
writing PHY page 769 (or 0x6020 shifted) reg 0x14
Later the link went down again with dynamic debug enabled. It came back
22 seconds later:
[ 2848.096035] [2031] e1000e:__e1000_read_phy_reg_hv:2839: e1000e
0000:00:1f.6 net00: reading PHY page 0 (or 0x0 shifted) reg 0x1
[ 2848.096121] [2031] e1000e:__e1000_read_phy_reg_hv:2839: e1000e
0000:00:1f.6 net00: reading PHY page 0 (or 0x0 shifted) reg 0x1
[ 2848.096200] e1000e 0000:00:1f.6 net00: NIC Link is Down
[…]
[ 2906.442227] [2031] e1000e:__e1000_read_phy_reg_hv:2839: e1000e
0000:00:1f.6 net00: reading PHY page 0 (or 0x0 shifted) reg 0xf
[ 2906.442299] [2031]
e1000e:e1000e_get_speed_and_duplex_copper:1321: e1000e 0000:00:1f.6
net00: 1000 Mbps, Full Duplex
[ 2906.442308] e1000e 0000:00:1f.6 net00: NIC Link is Up 1000 Mbps
Full Duplex, Flow Control: Rx/Tx
Can you deduce anything from the logs? Of course it could be floor
connector or the cable. I don’t want to change too much in the setup
though.
Kind regards,
Paul
PS: Output of `ethtool`:
```
@sighup:~$ sudo ethtool net00
Settings for net00:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supported pause frame use: Symmetric Receive-only
Supports auto-negotiation: Yes
Supported FEC modes: Not reported
Advertised link modes: 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Advertised FEC modes: Not reported
Link partner advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Link partner advertised pause frame use: Symmetric Receive-only
Link partner advertised auto-negotiation: Yes
Link partner advertised FEC modes: Not reported
Speed: 1000Mb/s
Duplex: Full
Auto-negotiation: on
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
MDI-X: on (auto)
Supports Wake-on: pumbg
Wake-on: g
Current message level: 0x00000007 (7)
drv probe link
Link detected: yes
```
Hi Paul,
Thanks for reporting this.
I went over the logs you shared and I didn't see anything obviously
wrong there. I suspect that it might be related to the manageability
engine (ME). I would like to suggest to try disabling ME or AMT it they
it is possible and see if it helps. You can also try removing all the
mei modules from the kernel.
I will try to deep dive into the logs in the upcoming days.
