Re: [PATCH net 1/2] ovs/vxlan: fix rtnl notifications on iface deletion

2016-06-13 Thread Nicolas Dichtel
Le 10/06/2016 19:52, Jesse Gross a écrit : [snip] > Does Geneve need this as well? > Yes, will fix it. Thank you, Nicolas

Re: [PATCH net 1/2] ovs/vxlan: fix rtnl notifications on iface deletion

2016-06-13 Thread Nicolas Dichtel
Le 10/06/2016 19:50, pravin shelar a écrit : [snip] > Patch looks good except the error handling code. earlier call to > vxlan_dev_configure() registers the net-device and add the device to > vxlan device list. it needs to be undone before freeing the device. > Ok, will fix it (and also for GRE).

Re: [PATCH net 1/2] ovs/vxlan: fix rtnl notifications on iface deletion

2016-06-10 Thread Jesse Gross
On Fri, Jun 10, 2016 at 2:32 AM, Nicolas Dichtel wrote: > The function vxlan_dev_create() (only used by ovs) never calls > rtnl_configure_link(). The consequence is that dev->rtnl_link_state is > never set to RTNL_LINK_INITIALIZED. > During the deletion phase, the function rollback_registered_many

Re: [PATCH net 1/2] ovs/vxlan: fix rtnl notifications on iface deletion

2016-06-10 Thread pravin shelar
On Fri, Jun 10, 2016 at 2:32 AM, Nicolas Dichtel wrote: > The function vxlan_dev_create() (only used by ovs) never calls > rtnl_configure_link(). The consequence is that dev->rtnl_link_state is > never set to RTNL_LINK_INITIALIZED. > During the deletion phase, the function rollback_registered_many

[PATCH net 1/2] ovs/vxlan: fix rtnl notifications on iface deletion

2016-06-10 Thread Nicolas Dichtel
The function vxlan_dev_create() (only used by ovs) never calls rtnl_configure_link(). The consequence is that dev->rtnl_link_state is never set to RTNL_LINK_INITIALIZED. During the deletion phase, the function rollback_registered_many() sends a RTM_DELLINK only if dev->rtnl_link_state is set to RTN