On (10/27/08 15:39), Peter Memishian wrote:
> > > http://zhadum.east.sun.com/ws/clearview/clearview-ipmpdev/webrev/
> > >
> > > Sowmini, please have a look.
> >
> > Looks good.
>
> Thanks, but the complexity in in.mpathd was bothering me, so I've tried
> another approach: have the caller indicate whether the IP routing-related
> tables should include IRE_MARK_TESTHIDDEN entries. The only catch with
> this approach is that there's no obvious place where the caller can
> request this. After examining all of the options, I've decided to
> introduce a "special" level the caller can set to request this information
> (EXPER_IP_AND_TESTHIDDEN). My hope is that this would be a short-term
> solution until we have a better MIB mechanism that allows for more
> selective retrieval.
>
> Anyway, it certainly makes the implementation simpler.
> The webrev has been respun.
>
Is there a webrev that I need to look at?
BTW, a question about nce_xmit:
2274 if (!IN6_IS_ADDR_UNSPECIFIED(sender) && type != ND_NEIGHBOR_ADVERT) {
2275 if ((ipif = ip_ndp_lookup_addr_v6(sender, ill)) == NULL)
:
2279 }
:
2295 if (IN6_IS_ADDR_UNSPECIFIED(sender) && !(flag & NDP_PROBE)) {
:
2313 } else if (!(IN6_IS_ADDR_UNSPECIFIED(sender))) {
2314 if ((src_ipif = ip_ndp_lookup_addr_v6(sender, ill)) == NULL ||
:
So for NA's we end up doing the ip_ndp_lookup_addr_v6() at line 2314
anyway, and for other packets (e.g., NS with !unspec sender) we would
end up doing the lookup twice. Why can't we just do one lookup at
line 2275 and track the ipif, since we need it for the zoneid for the
NA anyway?
--Sowmini
