Hi Guilherme,

First, the physical Mellanox card is on @zabaj1's server, so they would
need to do the final test on hardware.

I checked bug 2139598 (PR 569) in 0.107.1. That fixed udev renaming
interfaces, but it doesn't touch `sriov_drivers_autoprobe`.

The problem here is that Netplan writes `sriov_numvfs` while autoprobe
is still enabled. The kernel immediately tries to load `mlx5_core` on
each VF while Netplan tries to switch devlink to `switchdev`. The
Mellanox firmware can't handle both at once and times out with error
-110.

The fix in Netplan is simple: set `sriov_drivers_autoprobe` to 0 before
writing `sriov_numvfs` when switchdev is used. That way devlink switches
cleanly without fighting the kernel driver, and `netplan-sriov-rebind`
binds the VFs afterwards.

What do you think?

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2166228

Title:
  Mellanox NIC is unable to enter switchdev mode

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/netplan.io/+bug/2166228/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to