pimd: Tracking of RPF is *separate* from the lookup

Start the separation of tracking a Destination from the act
of looking it up.  The cojoining of these two concepts led
to a bunch of code that had to think about both problems leading
to weird situations and code paths.  Simplify the code by making
pim_ecmp_nexthop_search a static function and we only ever
call pim_ecmp_nexthop_lookup when we need to do a RPF().

pim_ecmp_nexthop_lookup will now attempt to find a stored pnc
and if it finds one it will report on the answer from it.

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

1 files changed, 4 insertions, 14 deletions

diff --git a/pimd/pim_rpf.c b/pimd/pim_rpf.c
index 55f788b5bb..8cc8b7481c 100644
--- a/pimd/pim_rpf.c
+++ b/pimd/pim_rpf.c
@@ -201,7 +201,6 @@ enum pim_rpf_result pim_rpf_update(struct pim_instance *pim,
 	struct pim_rpf *rpf = &up->rpf;
 	struct pim_rpf saved;
 	struct prefix nht_p;
-	struct pim_nexthop_cache pnc;
 	struct prefix src, grp;
 	bool neigh_needed = true;
 
@@ -232,23 +231,14 @@ enum pim_rpf_result pim_rpf_update(struct pim_instance *pim,
 	grp.family = AF_INET;
 	grp.prefixlen = IPV4_MAX_BITLEN;
 	grp.u.prefix4 = up->sg.grp;
-	memset(&pnc, 0, sizeof(struct pim_nexthop_cache));
 
 	if ((up->sg.src.s_addr == INADDR_ANY && I_am_RP(pim, up->sg.grp)) ||
 	    PIM_UPSTREAM_FLAG_TEST_FHR(up->flags))
 		neigh_needed = FALSE;
-	if (pim_find_or_track_nexthop(pim, &nht_p, up, NULL, &pnc)) {
-		if (pnc.nexthop_num) {
-			if (!pim_ecmp_nexthop_search(pim, &pnc,
-						     &up->rpf.source_nexthop,
-						     &src, &grp, neigh_needed))
-				return PIM_RPF_FAILURE;
-		}
-	} else {
-		if (!pim_ecmp_nexthop_lookup(pim, &rpf->source_nexthop, &src,
-					     &grp, neigh_needed))
-			return PIM_RPF_FAILURE;
-	}
+	pim_find_or_track_nexthop(pim, &nht_p, up, NULL, NULL);
+	if (!pim_ecmp_nexthop_lookup(pim, &rpf->source_nexthop, &src, &grp,
+				     neigh_needed))
+		return PIM_RPF_FAILURE;
 
 	rpf->rpf_addr.family = AF_INET;
 	rpf->rpf_addr.u.prefix4 = pim_rpf_find_rpf_addr(up);