[ovs-dev] [PATCH] openflow: Add OFPC_PORT_BLOCKED to ofp_capabilities

Ajinkya D Kadam ajinkya.kadam at nyu.edu
Sat Nov 26 08:08:57 UTC 2016


In case I need some clarifications should I ask it on this thread or create
a new thread on  discuss mailing list ?
ᐧ

On Sat, Nov 26, 2016 at 1:42 AM, Ajinkya D Kadam <ajinkya.kadam at nyu.edu>
wrote:

> Hi Ben,
>
> Thanks for your valuable input.
>
> I will work in this direction.
>
> On Fri, Nov 25, 2016 at 4:57 PM, Ben Pfaff <blp at ovn.org> wrote:
>
>> On Thu, Nov 24, 2016 at 05:35:29AM -0500, Ajinkya D Kadam wrote:
>> > This patch adds OFPC_PORT_BLOCKED capability to switch features which
>> > indicates that protocols outside of OpenFLow, such as Spanning Tree
>> (802.1D)
>> > will detect topology loops and block ports to prevent packet loops.
>> >
>> > If this bit is not set, then its the responsibility of the controller to
>> > implement a
>> > mechanism to prevent packet loops.
>> >
>> > CC: Ben Pfaff blp at ovn.org
>> > Signed-off-by: Ajinkya Kadam <ajinkya.kadam at nyu.edu>
>> >
>> > ---
>> >  include/openflow/openflow-common.h | 3 ++-
>> >  1 file changed, 2 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/include/openflow/openflow-common.h
>> > b/include/openflow/openflow-common.h
>> > index 7b619a9..492bd65 100644
>> > --- a/include/openflow/openflow-common.h
>> > +++ b/include/openflow/openflow-common.h
>> > @@ -276,8 +276,9 @@ enum ofp_capabilities {
>> >      OFPC_PORT_STATS     = 1 << 2,  /* Port statistics. */
>> >      OFPC_IP_REASM       = 1 << 5,  /* Can reassemble IP fragments. */
>> >      OFPC_QUEUE_STATS    = 1 << 6,  /* Queue statistics. */
>> > -    OFPC_ARP_MATCH_IP   = 1 << 7   /* Match IP addresses in ARP
>> > +    OFPC_ARP_MATCH_IP   = 1 << 7,  /* Match IP addresses in ARP
>> >                                        pkts. */
>> > +    OFPC_PORT_BLOCKED   = 1 << 8   /* Switch will block looping ports.
>> */
>> >  };
>>
>> I don't think the actual feature behind this bit is implemented.  That's
>> the part that would be valuable to implement.  Just adding the enum
>> value doesn't do much for us, especially since (as Simon points out),
>> it's already available with another name.
>>
>
>>


More information about the dev mailing list