On Wed, Jun 07, 2023 at 12:21:08PM +0200, David Marchand wrote: > Hello, > > On Tue, Jun 6, 2023 at 7:10 AM Eli Britstein <[email protected]> wrote: > > >> > > > >> > Vport's offloads are done on the tracked orig-in-port, but the flow > > >> > itself is associated in the vport's map. > > >> > > > >> > Removing the physdev will flush all the ports that are on its map, > > >> > but > > >> > > >> all the flows* > > >> > > >> > not the ones on other netdevs' maps. Since flows take reference > > >> > count on both their vport and their physdev, the physdev fails to be > > >removed. > > >> > > >> I tested with a simple ping over a vxlan tunnel. > > >> In my testing, I do manage to remove the physdev port. > > [snip] > > > >I am simply seeing the vport flow expiring. > > >So I am probably not testing the right way, could you share a reproducer? > > > > Sorry. Indeed, the commit msg is misleading. > > The physport is being able to be detached, while there is a tunnel > > offloaded traffic. > > However, re-adding it (before the flows are aged) encounters an "already in > > use" error: > > $ ovs-vsctl del-port p0 > > $ ovs-vsctl add-port br-phy0 p0 -- set interface p0 type=dpdk > > options:dpdk-devargs="0000:08:00.0,dv_xmeta_en=3" > > ovs-vsctl: Error detected while setting up 'p0': 'p0' is trying to use > > device '0000:08:00.0,dv_xmeta_en=3' which is already in use by 'p0'. See > > ovs-vswitchd log for details. > > ovs-vsctl: The default log directory is "/var/log/openvswitch". > > Ok, got it thanks. > It seems a bit strange to remove and re-add the PF port, but cleaning > up those flows makes sense on the principle. > > > > > I can rebase and rephrase the commit msg. > > Maybe rephrase yes, otherwise the fix works and lgtm.
Thanks, I think it would be best to post a v2 with this change. Please include patch 2/2 unless you wish to drop it for some reason. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
