Ethtool didn't support link speed 25Gb 56Gb 100Gb 200Gb

[Talat Batheesh]

please add missing Advertised speeds into ethtool

Ethtool 3.13 didn't support link speed 25Gb 56Gb 100Gb 200Gb.

Ethtool 4.2 didn't support link speed 25Gb 100Gb 200Gb.

This the list from ethtool 4.2
#define ALL_ADVERTISED_MODES \
        (ADVERTISED_10baseT_Half | \
         ADVERTISED_10baseT_Full | \
         ADVERTISED_100baseT_Half | \
         ADVERTISED_100baseT_Full | \
         ADVERTISED_1000baseT_Half | \
         ADVERTISED_1000baseT_Full | \
         ADVERTISED_1000baseKX_Full| \
         ADVERTISED_2500baseX_Full | \
         ADVERTISED_10000baseT_Full | \
         ADVERTISED_10000baseKX4_Full | \
         ADVERTISED_10000baseKR_Full | \
         ADVERTISED_10000baseR_FEC | \
         ADVERTISED_20000baseMLD2_Full | \
         ADVERTISED_20000baseKR2_Full | \
         ADVERTISED_40000baseKR4_Full | \
         ADVERTISED_40000baseCR4_Full | \
         ADVERTISED_40000baseSR4_Full | \
         ADVERTISED_40000baseLR4_Full | \
         ADVERTISED_56000baseKR4_Full | \
         ADVERTISED_56000baseCR4_Full | \
         ADVERTISED_56000baseSR4_Full | \
         ADVERTISED_56000baseLR4_Full)

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 1533721

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Joseph Salisbury (jsalisbury)]

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v4.4 kernel[0].

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

Once testing of the upstream kernel is complete, please mark this bug as "Confirmed".

Thanks in advance.

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.4-wily

[Ben Hutchings (benh-debian)]

Those are all the advertised flags defined by the kernel. This does not prevent forcing higher link speeds, only enabling auto-negotiation of those speeds.

[Rafael David Tinoco (rafaeldtinoco)]

After talking to Talat:

- What Ubuntu version are you referring to ?
[TB]Ubuntu 14.4.4 trusty

- Have you tested in newer versions ?
[TB] not tested yet, will test it on Xenial and update.

- What are the steps to reproduce the problem ?
[TB] ethtool <interface>
 See that under "Advertised link modes" not mentioned all the speeds that supported on this network interface
 ConnectX4 device support several speeds line 10 25 40 50 100
 With trusty version we see just
         Supported link modes: 10000baseKR/Full
                               40000baseCR4/Full
                               40000baseSR4/Full
                               40000baseLR4/Full

- Is there a known fix you are aware of ? Where ?
[TB] yes there are a fix for 56 will search for the commit and send later.

- Is Mellanox responsible for the code being fixed ? Or for the commit(s) fixing the issue ?
[TB] Yes it's fixed, will ask the maintainer of this and update

I'm coming back to this after some other IB bugs. So, Talat, if you could provide the commits fixing the issue I would appreciate.

Thank you

Rafael

[Talat Batheesh (talat-b87)]

Hi,

unfortunately, ethtool header does not currently support link speeds 25/50/100/200,
 therefore these speeds cannot be seen as dvertised/supported.

See this artical that will also expain the issue and suggest a fix.
https://lwn.net/Articles/672517/

This commit is an example of adding speeds to ethtool
b55a631d5ace86517508151dda491ffd40ad7e4b
http://lists.alioth.debian.org/pipermail/kernel-svn-changes/2015-October/021077.html

[Jeff Lane  (bladernr)]

To fix this, update ethtool in Ubuntu to the later versions. I have run into this issue now in Cert when customers are testing higher end NIC devices. The current version in Xenial is 4.5, and it looks like the current version in Artful is 4.8.

My customer tested this with 4.11 and you can see the differences below:
ethtool 1:4.5-1
Settings for enp94s0:
 Supported ports: [ FIBRE Backplane ]
 Supported link modes: 1000baseKX/Full
                         10000baseKR/Full
                         40000baseKR4/Full
                         40000baseCR4/Full
                         40000baseSR4/Full
                         40000baseLR4/Full
 Supported pause frame use: Symmetric Receive-only
 Supports auto-negotiation: Yes
 Advertised link modes: 1000baseKX/Full
                         10000baseKR/Full
                         40000baseKR4/Full
                         40000baseCR4/Full
                         40000baseSR4/Full
                         40000baseLR4/Full
 Advertised pause frame use: No
 Advertised auto-negotiation: Yes
 Link partner advertised link modes: Not reported
 Link partner advertised pause frame use: No
 Link partner advertised auto-negotiation: Yes
 Speed: 100000Mb/s
 Duplex: Full
 Port: Direct Attach Copper
 PHYAD: 0
 Transceiver: internal
 Auto-negotiation: on
 Supports Wake-on: d
 Wake-on: d
 Current message level: 0x00000004 (4)
          link
 Link detected: yes

And upstream 4.11:
Settings for enp94s0:
 Supported ports: [ FIBRE Backplane ]
 Supported link modes: 1000baseKX/Full
                         10000baseKR/Full
                         40000baseKR4/Full
                         40000baseCR4/Full
                         40000baseSR4/Full
                         40000baseLR4/Full
                         25000baseCR/Full
                         25000baseKR/Full
                         25000baseSR/Full
                         50000baseCR2/Full
                         50000baseKR2/Full
                         100000baseKR4/Full
                         100000baseSR4/Full
                         100000baseCR4/Full
                         100000baseLR4_ER4/Full
 Supported pause frame use: Symmetric Receive-only
 Supports auto-negotiation: Yes
 Advertised link modes: 1000baseKX/Full
                         10000baseKR/Full
                         40000baseKR4/Full
                         40000baseCR4/Full
                         40000baseSR4/Full
                         40000baseLR4/Full
                         25000baseCR/Full
                         25000baseKR/Full
                         25000baseSR/Full
                         50000baseCR2/Full
                         50000baseKR2/Full
                         100000baseKR4/Full
                         100000baseSR4/Full
                         100000baseCR4/Full
                         100000baseLR4_ER4/Full
 Advertised pause frame use: No
 Advertised auto-negotiation: Yes
 Speed: 100000Mb/s
 Duplex: Full
 Port: FIBRE
 PHYAD: 0
 Transceiver: internal
 Auto-negotiation: on
 Supports Wake-on: d
 Wake-on: d
 Current message level: 0x00000004 (4)
          link
 Link detected: yes

Ple...

[Jeff Lane  (bladernr)]

This is over a year and a half old at this point. IS there a status update for this issue?

[Christian Ehrhardt  (paelzer)]

We provide newer ethtool with newer kernels in newer releases.
So e.g. 4.15 in Bionic along the kernel with the same number.

This also is the reason this is now owned by the kernel team.
But AFAIK there are no plans to provide backports of ethtool along the HWE kernels (SRUs would have wuite some regression risks, maybe via backports pocket).
IMHO that is what you would need (ethtool in backports along HWE), or a ethtool-HWE package that conflicts with the base ethtool.

[Christian Ehrhardt  (paelzer)]

Subscribing a few kernel people to decide on further activity on this.