[ovs-dev] [PATCH] ofproto: fix interactions between flow monitors and barriers

[YAMAMOTO Takashi]

> On Wed, Feb 05, 2014 at 07:04:05PM +0900, YAMAMOTO Takashi wrote:
>> Following OpenFlow 1.4 semantics, make barriers wait for
>> flow monitor replies.  This should fix a race in
>> "ofproto - flow monitoring pause and resume" test.
>> 
>> Signed-off-by: YAMAMOTO Takashi <yamamoto at valinux.co.jp>
> 
> I still don't understand this.  You cite the following:
> 
>      * OpenFlow 1.4.0 p.110:
>      *    A OFPMP_FLOW_MONITOR multipart reply can not cross a barrier
>      *    handshake.  The switch must always send the OFPMP_FLOW_MONITOR
>      *    multipart reply for a given flow table change before the reply
>      *    to a OFPT_BARRIER_REQUEST request that follows the request
>      *    responsible for the flow table change.
> 
> but can you elaborate on your thought process here?  After all, when
> we're paused we're not sending any flow table changes at all.

right.  and when resumed, we send flow table changes.
the "ofproto - flow monitoring pause and resume" test seems to
expect that it can wait for these flow table changes after resume
using a barrier.  the expectation seems valid from my reading of
the specification.

YAMAMOTO Takashi