Thanks, that makes sense.
On Mon, Jul 22, 2013 at 12:51 PM, Andy Zhou <[email protected]> wrote: > The actual case I found this bug is more complicated to explain, but the > following made up example can illustrate the problem. Consider we have 2 > kernel mega flows: > > 1) in_port(1/0xffff), * > 2) in_port(2/0xffff), src_ip(1.1.1.1/255.255.255.255) > > They are not overlapping flows, however, when the following packet arrives > (in_port(2), src_ip(2,2,2,2)). if we apply mask of 1), it could match flow > 2). But it should not. > > Adding the mask check would have prevented this false match. > > --andy > > > On Mon, Jul 22, 2013 at 9:57 AM, Jesse Gross <[email protected]> wrote: >> >> On Fri, Jul 19, 2013 at 11:11 AM, Andy Zhou <[email protected]> wrote: >> > A mega flow matches when the masked key matches and the mask applied >> > is the same as the mask used to create the mega flow. >> > >> > This patch adds the implementation of the second match condition >> > mentioned above. Without this fix, mega flow lookup may result false >> > match. >> > >> > Bug #18584 >> > >> > Signed-off-by: Andy Zhou <[email protected]> >> >> Can you explain the conditions under which this can occur? It seems to >> imply the overlapping flows are being installed, which is not supposed >> to happen. > > _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
