1 Gbps link / NetworkManager problem

Hi
Tumbleweed, Thinkpad T450s

<deano_ferrari>
Your solution has been very helpful.
It was 100 Mbps Half and after entering “options SmartPowerDownEnable = 1” it got 100 Mbps FDX

But I have one more problem.
There should be 1 Gbps FDX and now there is 100 Mbps FDX
If I set in NetworkManager:
“auto-negotiate = true” - I’m losing my connection
or
“speed = 1000” - I’m losing my connection

“speed=100” autoneg=off - link ok

Short CAT6e cable, switch / mikrotik 1 Gbps port
Configuration of the ethernet card below:

ethtool enp0s25
Settings for enp0s25:
Supported ports: TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supported pause frame use: No
Supports auto-negotiation: Yes
Supported FEC modes: Not reported
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Advertised pause frame use: No
Advertised auto-negotiation: Yes
Advertised FEC modes: Not reported
Speed: 100Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
MDI-X: on (auto)
Supports Wake-on: pumbg
Wake-on: d
Current message level: 0x00000007 (7)
drv probe link
Link detected: yes


dmesg | grep e1000e 
    8.840119] e1000e: Intel(R) PRO/1000 Network Driver - 3.2.6-k 
    8.840121] e1000e: Copyright(c) 1999 - 2015 Intel Corporation. 
    8.840328] e1000e 0000:00:19.0: Interrupt Throttling Rate (ints/sec)  set to dynamic conservative mode 
    9.019886] e1000e 0000:00:19.0 0000:00:19.0 (uninitialized):  registered PHC clock 
    9.142318] e1000e 0000:00:19.0 eth0: (PCI Express:2.5GT/s:Width x1)  50:7b:9d:xx:xx:xx 
    9.142321] e1000e 0000:00:19.0 eth0: Intel(R) PRO/1000 Network  Connection 
    9.142471] e1000e 0000:00:19.0 eth0: MAC: 11, PHY: 12, PBA No:  1000FF-0FF 
    9.190099] e1000e 0000:00:19.0 enp0s25: renamed from eth0 
   33.474832] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: None 
   33.474839] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO 
  186.938805] e1000e: enp0s25 NIC Link is Down 
  212.347391] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: None 
  212.347399] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO 
 1228.961414] e1000e: enp0s25 NIC Link is Down 
 1274.294397] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: Rx/Tx 
 1274.294405] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO 
 1293.595053] e1000e: enp0s25 NIC Link is Down 
 1323.659645] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: None 
 1323.659652] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO 
 1830.001717] e1000e: enp0s25 NIC Link is Down 
 1851.809154] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: None 
 1851.809162] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO 
 2031.457852] e1000e: enp0s25 NIC Link is Down 
 2041.836066] e1000e: enp0s25 NIC Link is Down 
 2725.295174] e1000e: enp0s25 NIC Link is Down 
 2790.116636] e1000e: enp0s25 NIC Link is Up 100 Mbps Full Duplex, Flow  Control: Rx/Tx 
 2790.116643] e1000e 0000:00:19.0 enp0s25: 10/100 speed: disabling TSO

ifconfig 
enp0s25: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500 
        ether 50:7b:9d:xx:xx:xx  txqueuelen 1000  (Ethernet) 
        RX packets 5296  bytes 4569818 (4.3 MiB) 
        RX errors 0  dropped 798  overruns 0  frame 0 
        TX packets 3482  bytes 388654 (379.5 KiB) 
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0 
        device interrupt 20  memory 0xe1200000-e1220000

Please post output between CODE tags, the # in the layout ribbon. Makes it more readable, certainly after a long day of work :smiley: . Will add them now, so you can see the difference.

FWIW: I’ve seen this on the wired connection of my laptop ( different NIC though ). Before I could dive into it my ISP put new firmware on my modem/router and rebooted it ( on my request ). Ever since I have a 1Gbps.
On some other machine ( PC ) that used Networkmanager, I switched it to wicked networking, and got 1Gbps, went back to Networkmanager and that worked as expected from then on, i.e. owner never complained again.

So question is: is this a PC? Then why use Networkmanager anyway ?

The device on the other side does not matter. I have several dozen network devices with a 1 Gbps port and on the old x201 thinkpad (leap) there is 1 Gbps. At the beginning of the t450s there was also 1 Gbps - later it disappeared and became 100 FDX.
I use NetManager because I have a lot of different connections (vlan, vpn, wifi …) and I need to switch quickly and conveniently.
Once the network configuration in opensuse was wonderful. Now network configuration in tumbleweed is a tragedy. I have been using opensuse since 2002 (I bought a CD box with books :slight_smile: suse 7? )
Maybe there is a problem with the ethernet card in this thinkpad because the 2nd partition has win7 and there is only 100 fdx there.

Sorry, got interrupted by a silly phone call. Edited the OP now.

Are you allowing slower speeds intentionally?

ethtool enp0s25
Settings for enp0s25:
Supported ports: TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supported pause frame use: No
Supports auto-negotiation:** Yes**
Supported FEC modes: Not reported
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Advertised pause frame use: No
Advertised auto-negotiation:** Yes**
Advertised FEC modes: Not reported
Speed: 100Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
MDI-X: on (auto)
Supports Wake-on: pumbg
Wake-on: d
Current message level: 0x00000007 (7)
drv probe link
Link detected: yes

TSU

Usual first suspect is cable. 1Gbps requires all 8 wires while 100Mbs works with 4 wires. So any damaged single wire will downgrade connection to 100Mbps.

That’s good advice.

Also, if none of the dynamic features of NetworkManager (changing Wifi while changing location, being mobile) and/or DHCP/Bonjour etc are relevant for your 1Gbit/s setup, consider systemd-networkd.

Over the years, I switched from wicked to NM (much quicker startup than wicked) and finally, half a year ago, to networkd and dedicated, static IP addresses in my LAN. Bootup times improved further, logging verbosity improved noticably as well (NM and wicked both are quite chatty). Since then, I haven’t spent any further thought on my LAN setup. It’s nice and simple, naturally integrates well with other systemd services, with a fraction of the footprint of NM or wicked. (A similar improvement using display managers was switching from sddm to gdm, then xdm, then unthemed kdm, really fast, especially if you auto-login to KDE/Plasma5).

(https://forums.opensuse.org/showthread.php/535043-et0-not-(auto)-enabled-at-boot-time-no-static-ip-addresses-are-assigned?p=2896953#post2896953) Karl Mistelberger then gave further info about an even easier switch to networkd with DHCP.

  1. Cables are not a problem, I used different, short cat6e cables. On these cables, the x201 thinpad gets 1 gbps. This was the first thing I checked.
  2. I use different connections to configure different 100/1000 devices + a lot of vlan. systemd-networkd will not work in this situation, too much fun. It must be fast and easy.
    Which usb3 ethernet card do you recommend? No problems with tumbleweed drivers and 1 gbps speed.

Just a quick comment.

I have a laptop, where network performance became very poor on NetworkManager. Switching to “wicked” fixed it.

But then I removed all NetworkManager configuration – everything in “/etc/NetworkManager/system-connections” and in “/var/lib/NetworkManager”. Then I switched back to NetworkManager and all was fine.

It looks as if the problem was continually upgrading, with NetworkManager having changed in a way that was incompatible with its earlier configuration. And, of course, Tumbleweed uses continual upgrading.