No need to provide a dummy nexthop anymore, instead set the nhflags to

NEXTHOP_VALID or 0 depending on the nexthop state for the testvector.

diff --git a/regress/usr.sbin/bgpd/unittests/rde_decide_test.c b/regress/usr.sbin/bgpd/unittests/rde_decide_test.c
index 9086519..6d643dd 100644 (file)
--- a/regress/usr.sbin/bgpd/unittests/rde_decide_test.c
+++ b/regress/usr.sbin/bgpd/unittests/rde_decide_test.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: rde_decide_test.c,v 1.11 2022/07/19 16:27:59 claudio Exp $ */
+/*     $OpenBSD: rde_decide_test.c,v 1.12 2022/07/26 06:48:18 claudio Exp $ */
 
 /*
  * Copyright (c) 2020 Claudio Jeker <claudio@openbsd.org>
@@ -34,13 +34,6 @@ struct rib dummy_rib = {
 
 struct rib_entry dummy_re;
 
-struct nexthop nh_reach = {
-       .state = NEXTHOP_REACH
-};
-struct nexthop nh_unreach = {
-       .state = NEXTHOP_UNREACH
-};
-
 struct rde_peer peer1 = {
        .conf.ebgp = 1,
        .remote_bgpid = 1,
@@ -110,47 +103,47 @@ struct test {
        struct prefix p;
 } test_pfx[] = {
        { .what = "test prefix",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* pathes with errors are not eligible */
        { .what = "prefix with error",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[1], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[1], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* only loop free pathes are eligible */
        { .what = "prefix with loop",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[2], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[2], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 1. check if prefix is eligible a.k.a reachable */
        { .what = "prefix with unreachable nexthop",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nexthop = &nh_unreach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nhflags = 0, .lastchange = T1, } },
        /* 2. local preference of prefix, bigger is better */
        { .what = "local preference check",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[3], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[3], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 3. aspath count, the shorter the better */
        { .what = "aspath count check",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[4], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[4], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 4. origin, the lower the better */
        { .what = "origin EGP",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[5], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[5], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        { .what = "origin INCOMPLETE",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[6], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[6], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 5. MED decision */
        { .what = "MED",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[7], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[7], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 6. EBGP is cooler than IBGP */
        { .what = "EBGP vs IBGP",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1_i, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1_i, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 7. weight */
        { .what = "local weight",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[8], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[8], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 8. nexthop cost not implemented */
        /* 9. route age */
        { .what = "route age",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T2, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T2, } },
        /* 10. BGP Id or ORIGINATOR_ID if present */
        { .what = "BGP ID",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer2, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer2, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
        /* 11. CLUSTER_LIST length, TODO */
        /* 12. lowest peer address wins */
        { .what = "remote peer address",
-       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1_a4, .nexthop = &nh_reach, .lastchange = T1, } },
+       .p = { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1_a4, .nhflags = NEXTHOP_VALID, .lastchange = T1, } },
 };
 
 struct rde_aspath med_asp[] = {
@@ -170,27 +163,27 @@ struct rde_aspath med_asp[] = {
  * For med_pfx3 and med_pfx4 the strict case differs in the bgp-id.
  */
 struct prefix med_pfx1 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[0], .peer = &peer2, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[0], .peer = &peer2, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 struct prefix med_pfx2 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[1], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[1], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 struct prefix med_pfx3 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[2], .peer = &peer3, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[2], .peer = &peer3, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 struct prefix med_pfx4 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[3], .peer = &peer1_a4, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[3], .peer = &peer1_a4, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 /* the next two prefixes have a longer aspath than med_pfx1 & 2 */
 struct prefix med_pfx5 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[5], .peer = &peer3, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[5], .peer = &peer3, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 struct prefix med_pfx6 = 
-       { .entry.list.re = &dummy_re, .aspath = &med_asp[4], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &med_asp[4], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 
 /*
  * Define two prefixes where pfx1 > pfx2 if 'rde route-age evaluate'
  * but pfx1 < pfx2 if 'rde route-age ignore' 
  */
 struct prefix age_pfx1 = 
-       { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer2, .nexthop = &nh_reach, .lastchange = T1, };
+       { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer2, .nhflags = NEXTHOP_VALID, .lastchange = T1, };
 struct prefix age_pfx2 = 
-       { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nexthop = &nh_reach, .lastchange = T2, };
+       { .entry.list.re = &dummy_re, .aspath = &asp[0], .peer = &peer1, .nhflags = NEXTHOP_VALID, .lastchange = T2, };
 
 int     prefix_cmp(struct prefix *, struct prefix *, int *);