Score:0

Intel Pro/1000 EB on-board interface runs at 100 Mb/s not 1 Gp/s

cn flag

This seems to be the same as another question asked yesterday.

I’m having a problem with getting an Intel gigabit interface to operate at full speed.

System is a Supermicro X7DA8 dual Xeon. It has two Intel “Pro/1000 EB” interfaces on the motherboard.

The system dual-boots Ubuntu Studio 20.04.4 and MS Windows 7 Pro x64.

The system is used to back up a Supermicro server with 8 4TB SAS disks. That replaced a much older quad Opteron server where the speed was not a problem, but the new server can saturate a GigE wire most of the time and has a much larger partition, so GigE speeds are necessary.

The interface is recognized as a 1 Gb/s interface by Windows 7; I previously ran FreeBSD on it (my server back end has been FreeBSD since the early 1990’s) and that always connected at 1 Gb/s.

The other question had a request to run ethtool:

ethtool enp6s0f1
Settings for enp6s0f1:
    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)
Cannot get wake-on-lan settings: Operation not permitted
    Current message level: 0x00000007 (7)
                   drv probe link
    Link detected: yes

I ran the usual lshw and dmesg and discovered in the dmesg that the following:

[52.221560] e1000e 0000:06:00.1 enp6s0f1: Link Speed was downgraded by SmartSpeed.

The driver code seems to choose “Interrupt Throttling Rate (ints/sec) set to dynamic conservative” if the rate will be 100 Mb/s, but I haven’t looked any further (yet).

which is a signal of a bad cable.

The system was connected with a new cable labeled “CAT 6” but I bought a new “CAT 6” cable and when I installed the cable link speed was 1 GB/ps. After a reboot it was 100 Mb/s, and disconnecting and reconnecting didn’t change that.

I tried booting the current Xubuntu version 22.04.2 and it also linked at 100 Mb/s.

I have seen a recent note that there was a kernel bug affecting the Intel chip but I haven’t seen the real announcement so far.

I’ve ordered a non-Intel GigE network card which may provide a workaround.

dmesg output filtered for e1000 is:

[    1.274549] e1000e: Intel(R) PRO/1000 Network Driver
[    1.274593] e1000e: Copyright(c) 1999 - 2015 Intel Corporation.
[    1.280921] e1000e 0000:06:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
[    1.341558] e1000e 0000:06:00.0 eth0: (PCI Express:2.5GT/s:Width x4) 00:30:48:7e:cb:cc
[    1.341577] e1000e 0000:06:00.0 eth0: Intel(R) PRO/1000 Network Connection
[    1.341658] e1000e 0000:06:00.0 eth0: MAC: 5, PHY: 5, PBA No: 2050FF-0FF
[    1.341918] e1000e 0000:06:00.1: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
[    1.382426] e1000e 0000:06:00.1 eth1: (PCI Express:2.5GT/s:Width x4) 00:30:48:7e:cb:cd
[    1.382431] e1000e 0000:06:00.1 eth1: Intel(R) PRO/1000 Network Connection
[    1.382511] e1000e 0000:06:00.1 eth1: MAC: 5, PHY: 5, PBA No: 2050FF-0FF
[    1.384825] e1000e 0000:06:00.1 enp6s0f1: renamed from eth1
[    1.399218] e1000e 0000:06:00.0 enp6s0f0: renamed from eth0
[   52.221436] e1000e 0000:06:00.1 enp6s0f1: NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx
[   52.221560] e1000e 0000:06:00.1 enp6s0f1: Link Speed was downgraded by SmartSpeed
[   52.221563] e1000e 0000:06:00.1 enp6s0f1: 10/100 speed: disabling TSO

sudo lshw -C net yields

 *-network:0
       description: Ethernet interface
       product: 80003ES2LAN Gigabit Ethernet Controller (Copper)
       vendor: Intel Corporation
       physical id: 0
       bus info: pci@0000:06:00.0
       logical name: enp6s0f0
       version: 01
       serial: 00:30:48:7e:cb:cc
       capacity: 1Gbit/s
       width: 32 bits
       clock: 33MHz
       capabilities: pm msi pciexpress bus_master cap_list rom ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=e1000e driverversion=5.15.0-70-lowlatency firmware=2.1-12 latency=0 link=no multicast=yes port=twisted pair
       resources: irq:24 memory:d2220000-d223ffff memory:d2200000-d221ffff ioport:4000(size=32) memory:d2280000-d228ffff
  *-network:1
       description: Ethernet interface
       product: 80003ES2LAN Gigabit Ethernet Controller (Copper)
       vendor: Intel Corporation
       physical id: 0.1
       bus info: pci@0000:06:00.1
       logical name: enp6s0f1
       version: 01
       serial: 00:30:48:7e:cb:cd
       size: 100Mbit/s
       capacity: 1Gbit/s
       width: 32 bits
       clock: 33MHz
       capabilities: pm msi pciexpress bus_master cap_list rom ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=e1000e driverversion=5.15.0-70-lowlatency duplex=full firmware=2.1-12 ip=192.168.0.25 latency=0 link=yes multicast=yes port=twisted pair speed=100Mbit/s
       resources: irq:26 memory:d2260000-d227ffff memory:d2240000-d225ffff ioport:4020(size=32) memory:d2290000-d229ffff
cc flag
There was a just reported bug in the e1000e driver, but it was for a 60% max, not 10% like you report. https://www.phoronix.com/forums/forum/software/general-linux-open-source/1383826-intel-i219-lm-had-only-been-running-at-~60-of-maximum-speed-due-to-linux-driver-bug
Mike Squires avatar
cn flag
Booted the system from the FreeBSD 13.1 DVD; active Ethernet interface connected at 1 Gb/s with no issues. Given that MS Windows 7 and FreeBSD 13.1 connect at 1 Gb/s but both Ubuntu Studio 20.04 and Xubuntu 22.04 at 100 Mb/s this would seem to suggest an Ubuntu issue. I will be trying to work around this with a non-Intel Ethernet card.
Mike Squires avatar
cn flag
Using "ethtool" I have reset the speed to 1 Gp/s full duplex; rsync is running at 15 Mb/s with a few peaks to 60 Mb/s which is more than before but much slower than rsync between the old FreeBSD server and the new one. The backup system is an Ubuntu Studio desktop which I can repurpose to FreeBSD if necessary but I'd have to bring up a new workstation. I have a generic Ethernet and a Broadcom Ethernet card on the way, so I will wait until those fail before repurposing.
Mike Squires avatar
cn flag
Problem solved by a $10 purchase of a Broadcomm PCI-X GigE Ethernet card. System now copying at 100 Mb/s with peaks to 200 Mb/s (buffering?). My suspicion is that the specific chip used on the Supermicro X7DAE board was never added to the driver. If I find the chip ID I'll ask again since it's a common chip in server/workstation motherboards.
cc flag
You may answer your own question, and if upvoted, mark it as solved -- gaining some rep. points and helping others with the problem.
I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.