Hi Yevgeny, On Mon, 2006-09-25 at 08:35, Yevgeny Kliteynik wrote: > Hi Hal. > > The patch looks ok.
Thanks. Did you just look at things or did you also run any tests ? > A few remarks thought: > > It appears that the multicast group mtu/rate selectors > are actually not referenced by anyone - the SM/SA code > implicitly assumes that they should be 'exact', and acts > accordingly. If this is the case, then it seems to me that it is a different bug which also needs fixing but I'm not sure why you say this as I see code in osm_sa_mcmember_record.c which obtains the various selectors (MTU and rate) and handles them. PacketLifeTime is the only one not currently handled. > Same goes for the response - the selectors > that are filled in are hard-coded to 'exact'. That's what the spec requires on response regardless of what selectors were supplied on the request. > This is the reason why the bug that this patch fixes has > never appeared, and why fixing it will not change the SM > behavior. It does change the SM (SA) behavior (at least in terms of the response). It does not affect the SA client (end stack) behavior. > But of course, it is better to have this fix anyway. Yes, I think it is required for spec compliance. -- Hal _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
