Alexander Bluhm([email protected]) on 2018.07.10 13:17:45 +0200: > On Tue, Jul 10, 2018 at 01:04:54PM +0200, Florian Riehm wrote: > > Several people, including myself, asked why we need the DELAY(1000) in > > netinet/ip_carp.c. It exists since the initial revision of carp(4). > > Nobody can exactly explain why it was added and tests work fine without it. > > I would like to remove it, since it blocks unlocking efforts of tb@ und > > mpi@. > > If contrary to expectations problems show up we should think about a better > > solution. > > Yes, commit it an see if some special setups break. This could > only happen if there are a lot of addresses on the carp interface. > > OK bluhm@
ok, (and for me its actually easier to test this when its commited). > > Index: netinet/ip_carp.c > > =================================================================== > > RCS file: /home/friehm/repos/openbsd-cvs/cvs/src/sys/netinet/ip_carp.c,v > > retrieving revision 1.332 > > diff -u -p -r1.332 ip_carp.c > > --- netinet/ip_carp.c 21 May 2018 15:52:22 -0000 1.332 > > +++ netinet/ip_carp.c 8 Jul 2018 16:19:57 -0000 > > @@ -1277,7 +1277,6 @@ carp_send_arp(struct carp_softc *sc) > > > > in = ifatoia(ifa)->ia_addr.sin_addr.s_addr; > > arprequest(&sc->sc_if, &in, &in, sc->sc_ac.ac_enaddr); > > - DELAY(1000); /* XXX */ > > } > > } > > > > @@ -1298,7 +1297,6 @@ carp_send_na(struct carp_softc *sc) > > nd6_na_output(&sc->sc_if, &mcast, in6, > > ND_NA_FLAG_OVERRIDE | > > (ip6_forwarding ? ND_NA_FLAG_ROUTER : 0), 1, NULL); > > - DELAY(1000); /* XXX */ > > } > > } > > #endif /* INET6 */ >
