PIM candidate BSR/RP support

[Jafaral]

This Brings full BSR and RP candidate support to FRR (rfc 5059)

Replaces #9180 , in coordination with @eqvinox

[donaldsharp]

Ran the test and got this new memory leaks:

  ## showing active allocations in memory group pimd
pimd:     PIM BSR C-RP                  :      2 *         96
pimd:     PIM BSR range                 :      3 *         80
pimd:     PIM BSR C-RP range item       :      4 *         88

My expectation is that we should not be adding new code that leaks

[Jafaral]

Ran the test and got this new memory leaks:

  ## showing active allocations in memory group pimd
pimd:     PIM BSR C-RP                  :      2 *         96
pimd:     PIM BSR range                 :      3 *         80
pimd:     PIM BSR C-RP range item       :      4 *         88

My expectation is that we should not be adding new code that leaks

will fix.

[nabahr]

Just a few minor comments

[mobash-rasool]

You can use "vty_json" instead of vty_out and json_object_free()

[donaldsharp]

Why was this resolved? I started to write a comment about this as well, I see that vty_json is not being used in several spots

[Jafaral]

Working on addressing comments from @nabahr and @mobash-rasool

[ton31337]

I suggest adding this also to JSON output, e.g.:

{"warning": "..."}

[Jafaral]

This isn't a warning really, just informational.

[Jafaral]

I.e, the correct behavior it return an empty object.

[ton31337]

Neither agree, nor disagree, e.g. we have in BGP such patterns:

	if (!peer) {
		if (use_json) {
			json_object *json_no = NULL;
			json_no = json_object_new_object();
			json_object_string_add(json_no, "warning",
					       "No such neighbor in this view/vrf");
			vty_json(vty, json_no);
		} else
			vty_out(vty, "No such neighbor in this view/vrf\n");
		return NULL;
	}

or

	if (is_synchronized(rpki_vrf)) {
		if (strmatch(prefixkind, "prefix-count"))
			print_prefix_table(vty, rpki_vrf, json, true);
		else
			print_prefix_table(vty, rpki_vrf, json, false);
	} else {
		if (json) {
			json_object_string_add(json, "error", "No Connection to RPKI cache server.");
			vty_json(vty, json);
		} else
			vty_out(vty, "No connection to RPKI cache server.\n");
		return CMD_WARNING;
	}

Anyway, it's not a blocker. Just an improvement.

[Jafaral]

Both of these are a bit different. The first , if you query a neighbor that doesn't exist, a warning is a good idea about it. The second case is an error, so that does need an error return. So I see why those have them. The case I have is purely informational.

[Jafaral]

I addressed all comments on this PR.