In Kernel 4.8 we introduced Traffic Control (TC see ) hardware offloads
framework for SR-IOV VFs which allows us to configure the NIC .
Subsequent OVS patches  allow us to use the TC framework
to offload OVS datapath rules. The feature is merged in ovs upstream.
To allow the support in openstack we need the following changes:
os-vif - https://review.openstack.org/#/c/460278/
This patch adds plug/unplug when using ovs vif 'OVS_ACCELERATION'. The
plug method will lookup the VF representor and connect it to 'br-int',
while the unplug method will remove the VF representor from 'br-int'.
This patch add support for VIF_TYPE_OVS_ACCELERATION
neutron patches - https://review.openstack.org/#/c/275616/ & https://review.openstack.org/#/c/452530/
This patch allow OVS mech driver to bind direct (SR-IOV) port.
This will allow to offload the OVS flows using tc to the SR-IOV NIC
and gain accelerate OVS.
The current support is for ovs mechanism driver we will extend the support to other mechanism drivers such odl and ovn.
This BZ depends on:
Bug 1467356 - [Mellanox Pike FEAT] openstack-neutron: Add support for OVS HW offload
Bug 1467354 - [Mellanox Pike FEAT] openstack-nova: add support for VIF_TYPE_OVS_ACCELERATION
Bug 1467220 - [Mellanox Pike FEAT] openvswitch: update package
Bug 1467359 - [Mellanox Pike FEAT] python-os-vif: update package
*** Bug 1468329 has been marked as a duplicate of this bug. ***
The content for this fix is included in OSP 13 RC which is available on the Red Hat OpenStack Platform beta channels currently. Any efforts to verify that these bugs have been fixed would be appreciated.
Do you mean to check ovs-offload works with OSP 13?
What exact bugs are you referring
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.