tests: Fix crash in test_mp_attr

Some recent changes in BGP now require that the
peer's nexthop have a valid ifp when we are looking
at:

	case BGP_ATTR_NHLEN_IPV6_GLOBAL_AND_LL:
	case BGP_ATTR_NHLEN_VPNV6_GLOBAL_AND_LL:

This assumption makes sense for this type of Nexthop Attribute.
So for the test let's jimmy up a `fake` enough interface pointer
so that the actual test we can focus on what we are actually
testing.

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

diff --git a/tests/bgpd/test_mp_attr.c b/tests/bgpd/test_mp_attr.c
index 34c35cfcc5872d543bc760331d74d5cb99f0848d..9e0cb1f5ca61cf3e7bb5af2e5bd621690f24324f 100644 (file)
--- a/tests/bgpd/test_mp_attr.c
+++ b/tests/bgpd/test_mp_attr.c
@@ -1033,6 +1033,7 @@ static as_t asn = 100;
 
 int main(void)
 {
+       struct interface ifp;
        struct peer *peer;
        int i, j;
 
@@ -1065,6 +1066,9 @@ int main(void)
        peer->status = Established;
        peer->curr = stream_new(BGP_MAX_PACKET_SIZE);
 
+       ifp.ifindex = 0;
+       peer->nexthop.ifp = &ifp;
+
        for (i = AFI_IP; i < AFI_MAX; i++)
                for (j = SAFI_UNICAST; j < SAFI_MAX; j++) {
                        peer->afc[i][j] = 1;