I created a ticket for this (EXT-315) to make sure we keep track of the issue, and fix/clarify it in the next revisions of the specs. Until then I suggest you do not use OFPP_MAX for ports, there's plenty of ids there anyway. :o)
Zoltan. >-----Original Message----- >From: Ben Pfaff [mailto:b...@nicira.com] >Sent: Tuesday, April 30, 2013 8:40 AM >To: Zoltán Lajos Kis >Cc: Bertrand Bonnefoy-Claudet; openflow-discuss@lists.stanford.edu >Subject: Re: [openflow-discuss] Port numbering and OFPP_MAX > >The OpenFlow 1.0 reference implementation code includes several uses of >OFPP_MAX. Most of them imply that a port number must be less than >OFPP_MAX. Two comments in OF1.0 openflow.h are explicit about this: > > uint16_t port; /* Port that queue belongs. Should > refer to a valid physical port > (i.e. < OFPP_MAX) or OFPP_IN_PORT. */ > > uint16_t port; /* Port to be queried. Should refer > to a valid physical port (i.e. < OFPP_MAX) */ > >One comment in OF1.3.1 openflow.h is also explicit about this: > > uint32_t port; /* Port to be queried. Should refer > to a valid physical port (i.e. < OFPP_MAX), > or OFPP_ANY to request all configured > queues.*/ > >On Tue, Apr 30, 2013 at 06:09:41AM +0000, Zolt?n Lajos Kis wrote: >> I think ports can go from 1 to OFPP_MAX inclusive, and the mistake is >> in the comment on page 88 (should be <= OFPP_MAX). >> >> Every xxx_MAX constant in OpenFlow refers to the last usable id >(OFPP_MAX, OFPG_MAX, OFPM_MAX, OFPTT_MAX, ...). In most cases the >structs explicitly say so. >> Ports are only special, because there the numbering starts from 1, which >makes the total number equal to the last usable id. >> >> Zoltan. >> >> >-----Original Message----- >> >From: openflow-discuss-boun...@lists.stanford.edu [mailto:openflow- >> >discuss-boun...@lists.stanford.edu] On Behalf Of Ben Pfaff >> >Sent: Tuesday, April 30, 2013 7:54 AM >> >To: Bertrand Bonnefoy-Claudet >> >Cc: openflow-discuss@lists.stanford.edu >> >Subject: Re: [openflow-discuss] Port numbering and OFPP_MAX >> > >> >On Mon, Apr 29, 2013 at 08:36:57PM -0500, Bertrand Bonnefoy-Claudet >wrote: >> >> Here is a minor point in the OpenFlow 1.3.1 specification (and most >> >> previous versions) that I would like to clarify. >> >> >> >> Page 42 says that ports are numbered starting from 1 and that >> >> OFPP_MAX is the maximum number of "normal" ports a switch can have. >> >> >> >> Page 88, on the other hand, says that a valid physical port should >> >> have a number "< OFPP_MAX". >> >> >> >> So, to my understanding, page 42 implies OFPP_MAX is a valid port >> >> whereas page 88 implies the contrary. >> > >> >This sounds like a mistake. I imagine that it arose because the >> >earliest versions of OpenFlow numbered ports starting from 0. The >> >intent is that port numbers are greater than 0 and less than OFPP_MAX. >> >_______________________________________________ >> >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
