Oops, probably mistakes due to retrieving it from my brain rather than working
code, but points (1) and (3) remain.
On Nov 7, 2011, at 12:15 PM, Bob Lantz wrote:
> Exactly - if everything else is wildcarded by default, why isn't DL_TYPE?
>
> In NOX, I seem to have to use something like { DL_TYPE: ethernet.IP_TYPE,
> NW_DST: '10.0.1.2' } for an L3 match.
>
> Moreover, if I'm telling the API to match an IPv4 address (e.g. {
> NW_DST:'10.0.1.2' } in NOX), wouldn't it be a reasonable default to assume
> that I'm going to want to match IPv4 packets - or at least *some* packets -
> rather than not matching anything at all?
>
> On Nov 7, 2011, at 12:03 PM, Josh Smift wrote:
>
>> BL> It seems that you should be able to simply specify L3 or L4 match
>> BL> fields via the NOX API, as Ramon is doing, and have everything both
>> BL> work on the switch *and* be reported correctly via dpctl, but that
>> BL> doesn't seem to be the case at this time.
>>
>> NB> I don't see how you could deterministically provide this "magic".
>> NB>
>> NB> - Lets say someone specifies nw_src - did they want the dl_type to be
>> NB> 0x800 or 0x806? nw_src is a valid field for both ethertypes.
>>
>> If they don't specify dl_type, they presumably want both (and any other
>> ethertype too), just like if they leave any other field blank.
>>
>> NB> - How about tp_src and tp_dst? What ip_proto should we assume that you
>> NB> are using?
>>
>> Does it need to assume that you're using any ip_proto?
>>
>> -Josh (j...@bbn.com)
>> _______________________________________________
>> openflow-discuss mailing list
>> openflow-discuss@lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
>
_______________________________________________
openflow-discuss mailing list
openflow-discuss@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
