On 13 June 2017 at 08:03, Roi Dayan <[email protected]> wrote: > From: Paul Blakey <[email protected]> > > To use netdev flow offloading api, dpifs needs to iterate over > added ports. This addition inserts the added dpif ports in a hash map, > The map will also be used to translate dpif ports to netdevs. > > Signed-off-by: Paul Blakey <[email protected]> > Reviewed-by: Roi Dayan <[email protected]> > Reviewed-by: Simon Horman <[email protected]> > Acked-by: Flavio Leitner <[email protected]>
Hi Roi, Simon, It looks like this patch has introduced a regression with regards to clean shutdown of OVS. Unfortunately I wasn't able to discover this until now because the tree has been a bit unstable lately with respect to running the system-traffic tests and OVS cleaning up after itself. I've bisected it down now, though. To observe the issue, run: # make check-system-userspace TESTSUITEFLAGS="1" # ip link show ovs-netdev If you find that the ovs-netdev device exists, then you are observing the problem. Typically when the 'check-system-userspace' tests run successfully, they should initiate complete and proper shutdown of OVS and removal of related state via "ovs-appctl exit --cleanup", which you can find invoked via the shell function kill_ovs_vswitchd which is queued up in _OVS_VSWITCHD_START via the OVS "on_exit" function. However, since this patch was merged the test no longer cleans up properly after itself. Note that if you run any subsequent tests without cleaning up, then they will fail. For instance, if you run tests 1-5 you'll see that 1 succeeds and 2-5 fail. Subsequent runs of any test will fail until you have removed the ovs-netdev device. Would you have some time to look into why ovs-vswitchd is not cleaning up this device during "ovs-appctl exit --cleanup" after this patch was applied? Thanks, Joe _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
