pimd, pim6d: route-map filtering for source/group

[rzalamena]

Implement route-map support for PIM / PIMv6 to allow users to filter IGMP / MLD joins using source / group / interface combinations.

[csiltala]

Looks good to me beyond the topotest failure, looks like there may be a couple places the previous multicast boundary wasn't being checked but should have been but that's a separate issue from the route-map functionality in this PR.

[Jafaral]

@rzalamena , one minor nit from frrbot, otherwise this is good to go.

--- a/pimd/pim6_mld.c
+++ b/pimd/pim6_mld.c
@@ -268,7 +268,8 @@ static bool gm_sg_limit_reached(struct gm_if *gm_if, const pim_addr source, cons
 	return false;
 }
 
-static bool gm_sg_filter_match(const struct gm_if *gm_if, const pim_addr source, const pim_addr group)
+static bool gm_sg_filter_match(const struct gm_if *gm_if, const pim_addr source,
+			       const pim_addr group)
 {

[rzalamena]

Update: the new topotest is failing on i386 consistently due to python magic in mcast-tester.py. I'm trying to figure out why it is not possible to do source specific join.

  File "/root/share/topotests/lib/mcast-tester.py", line 182, in <module>
    multicast_join(msock, ifindex, args.group, args.port, args.source)
  File "/root/share/topotests/lib/mcast-tester.py", line 114, in multicast_join
    sock.setsockopt(ip_proto, opt, mreq)
OSError: [Errno 99] Cannot assign requested address

[rzalamena]

I fixed the multicast tester issue and now the test is passing on i386.

[Jafaral]

I fixed the multicast tester issue and now the test is passing on i386.

Thanks @rzalamena . I will merge this PR, but do we need to backport the fix to 10.4?

[rzalamena]

I fixed the multicast tester issue and now the test is passing on i386.

Thanks @rzalamena . I will merge this PR, but do we need to backport the fix to 10.4?

No, this fix is for this new test using IPv6 MLD SG join. Previously nobody else was doing IPv6 MLD join with source and group using mcast-tester.py.