On Fri, Oct 14, 2016 at 7:12 AM, Thadeu Lima de Souza Cascardo
<casca...@redhat.com> wrote:
> There were leaks on netdev_close references. Whenever an address was removed
> from an interface, it was possible that a race would make the netlink message 
> be
> processed and no address be found on the interface, causing insert_ipdev to 
> bail
> out.
>
> When that happened, the netdev would be left opened as type system, and when
> adding an interface with the same name as type internal, even though the 
> system
> interface was already gone, netdev_open would return the one with type system.
> And, then, no internal interface would be created.
>
> Tested-by: Aaron Conole <acon...@redhat.com>
> Signed-off-by: Thadeu Lima de Souza Cascardo <casca...@redhat.com>
> ---
>
> So, Aaron found out that applying the following commits would fix a bug in
> branch 2.5, but we considered that would be too much. So, after some
> investigation, we realized that the change below fixes the issue, as described
> above. So, we would like to see this applied to branch 2.5, instead of doing a
> backport for a larger change.
>
> 2b02db1 packets: Add new functions for IPv4 and IPv6 address parsing.
> a8704b5 tunneling: Handle multiple ip address for given device.
> 3e6dc8b netdev: Verify ifa_addr is not NULL when iterating over getifaddrs.
> c2a1cee route-table: flush addresses list when route table is reset
>

Thanks for the patch. I pushed patch to branch 2.5.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to