[Bug 1948666] [NEW] Misleading behavior when enabling hardware offloads

[Alin-Gabriel Serdean]

Public bug reported:

According to the documentation(1) the following setting:
ovs-vsctl set Open_vSwitch . other_config:hw-offload=true|false/unset
should have effects only after vswitchd is restarted.

However, flipping the value on will have the following effect:
Hardware offloads will be enabled for newly created ports,
but not on the ports prior enabling the TC offloads.

If hw-offload got enabled, qdisc configured, hw-offload disabled --> ovs restarted.
This sequence might not delete the qdisc, and vswitchd on restart might not touch existing ones.
As a side effect you will have to delete them manually.

(1)
https://github.com/openvswitch/ovs/blob/v2.16.1/vswitchd/vswitch.xml#L225-L248

** Affects: openvswitch (Ubuntu)
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Ubuntu
OpenStack, which is subscribed to openvswitch in Ubuntu.
https://bugs.launchpad.net/bugs/1948666

Title:
  Misleading behavior when enabling hardware offloads

Status in openvswitch package in Ubuntu:
  New

Bug description:
  According to the documentation(1) the following setting:
  ovs-vsctl set Open_vSwitch . other_config:hw-offload=true|false/unset
  should have effects only after vswitchd is restarted.

  However, flipping the value on will have the following effect:
  Hardware offloads will be enabled for newly created ports,
  but not on the ports prior enabling the TC offloads.

  If hw-offload got enabled, qdisc configured, hw-offload disabled --> ovs restarted.
  This sequence might not delete the qdisc, and vswitchd on restart might not touch existing ones.
  As a side effect you will have to delete them manually.

  (1)
  https://github.com/openvswitch/ovs/blob/v2.16.1/vswitchd/vswitch.xml#L225-L248

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/openvswitch/+bug/1948666/+subscriptions