Commit 9458fc7041 for frr
commit 9458fc7041cc65947085446fc2f70bcf45e853b9
Author: Sougata Barik <sougatab@nvidia.com>
Date: Wed Nov 4 11:40:09 2026 +0530
tests: cover BFD after moving a BGP neighbor between peer-groups
r1 has peer-groups EDGE1/EDGE2 (multihop eBGP over loopbacks) and
SH1/SH2 (single-hop), all with BFD profile P1; r2 advertises 20000
sharpd routes so the old peer object lingers while its routes are
cleared. The test moves the neighbors EDGE1 -> EDGE2 / SH1 -> SH2,
moves them back, then deletes and re-adds them into EDGE2/SH2. After
each step bfdd on both routers must have the sessions Up and bgpd
must report BFD Up, also after the old peer objects are freed.
Without the fix r1's bfdd loses the 10.254.0.2 session after a move
while bgpd keeps showing BFD Up.
After each step the test also checks that r2's bgpd reports BFD Up and
that the neighbor is in the expected peer-group. Before the final check
it waits until r1's "BGP peer" memory count is back to its value before
the change, so the old peer object has really been freed.
Ticket: #5368914
Signed-off-by: Sougata Barik <sougatab@nvidia.com>
diff --git a/tests/topotests/bgp_bfd_peer_group_move/__init__.py b/tests/topotests/bgp_bfd_peer_group_move/__init__.py
new file mode 100644
index 0000000000..e69de29bb2
diff --git a/tests/topotests/bgp_bfd_peer_group_move/r1/frr.conf b/tests/topotests/bgp_bfd_peer_group_move/r1/frr.conf
new file mode 100644
index 0000000000..17ff5e2657
--- /dev/null
+++ b/tests/topotests/bgp_bfd_peer_group_move/r1/frr.conf
@@ -0,0 +1,47 @@
+!
+interface lo
+ ip address 10.254.0.1/32
+!
+interface r1-eth0
+ ip address 10.0.12.1/24
+!
+ip route 10.254.0.2/32 10.0.12.2
+!
+bfd
+ profile P1
+ receive-interval 300
+ transmit-interval 300
+ !
+!
+router bgp 65001
+ no bgp ebgp-requires-policy
+ no bgp network import-check
+ bgp router-id 10.254.0.1
+ neighbor EDGE1 peer-group
+ neighbor EDGE1 remote-as external
+ neighbor EDGE1 ebgp-multihop 2
+ neighbor EDGE1 update-source lo
+ neighbor EDGE1 bfd profile P1
+ neighbor EDGE1 timers 3 10
+ neighbor EDGE1 timers connect 1
+ neighbor EDGE2 peer-group
+ neighbor EDGE2 remote-as external
+ neighbor EDGE2 ebgp-multihop 2
+ neighbor EDGE2 update-source lo
+ neighbor EDGE2 bfd profile P1
+ neighbor EDGE2 timers 3 10
+ neighbor EDGE2 timers connect 1
+ neighbor SH1 peer-group
+ neighbor SH1 remote-as external
+ neighbor SH1 bfd profile P1
+ neighbor SH1 timers 3 10
+ neighbor SH1 timers connect 1
+ neighbor SH2 peer-group
+ neighbor SH2 remote-as external
+ neighbor SH2 bfd profile P1
+ neighbor SH2 timers 3 10
+ neighbor SH2 timers connect 1
+ neighbor 10.254.0.2 peer-group EDGE1
+ neighbor 10.0.12.2 peer-group SH1
+exit
+!
diff --git a/tests/topotests/bgp_bfd_peer_group_move/r2/frr.conf b/tests/topotests/bgp_bfd_peer_group_move/r2/frr.conf
new file mode 100644
index 0000000000..39f3771827
--- /dev/null
+++ b/tests/topotests/bgp_bfd_peer_group_move/r2/frr.conf
@@ -0,0 +1,35 @@
+!
+interface lo
+ ip address 10.254.0.2/32
+!
+interface r2-eth0
+ ip address 10.0.12.2/24
+!
+ip route 10.254.0.1/32 10.0.12.1
+!
+bfd
+ profile P1
+ receive-interval 300
+ transmit-interval 300
+ !
+!
+router bgp 65002
+ no bgp ebgp-requires-policy
+ bgp router-id 10.254.0.2
+ neighbor 10.254.0.1 remote-as external
+ neighbor 10.254.0.1 ebgp-multihop 2
+ neighbor 10.254.0.1 update-source lo
+ neighbor 10.254.0.1 bfd profile P1
+ neighbor 10.254.0.1 timers 3 10
+ neighbor 10.254.0.1 timers connect 1
+ neighbor 10.0.12.1 remote-as external
+ neighbor 10.0.12.1 bfd profile P1
+ neighbor 10.0.12.1 timers 3 10
+ neighbor 10.0.12.1 timers connect 1
+ address-family ipv4 unicast
+ redistribute sharp
+ neighbor 10.254.0.1 next-hop-self
+ neighbor 10.0.12.1 next-hop-self
+ exit-address-family
+exit
+!
diff --git a/tests/topotests/bgp_bfd_peer_group_move/test_bgp_bfd_peer_group_move.py b/tests/topotests/bgp_bfd_peer_group_move/test_bgp_bfd_peer_group_move.py
new file mode 100644
index 0000000000..4bffd83276
--- /dev/null
+++ b/tests/topotests/bgp_bfd_peer_group_move/test_bgp_bfd_peer_group_move.py
@@ -0,0 +1,246 @@
+#!/usr/bin/env python
+# SPDX-License-Identifier: ISC
+
+#
+# Copyright (c) 2026 by NVIDIA Corporation
+# Sougata Barik <sougatab@nvidia.com>
+#
+
+"""
+Move BGP neighbors that inherit BFD between peer-groups with the same BFD
+profile, while they carry enough routes for the old peer to linger in the
+Deleted state:
+
+ - multihop eBGP over loopbacks (EDGE1 <-> EDGE2)
+ - single-hop eBGP (SH1 <-> SH2)
+
+After every move, and after deleting and re-adding a neighbor, bfdd on both
+routers must still have the session Up and bgpd must agree with it.
+"""
+
+import os
+import re
+import sys
+import json
+import time
+import pytest
+import functools
+
+CWD = os.path.dirname(os.path.realpath(__file__))
+sys.path.append(os.path.join(CWD, "../"))
+
+# pylint: disable=C0413
+from lib import topotest
+from lib.topogen import Topogen, get_topogen, TopoRouter
+from lib.common_config import step
+
+pytestmark = [pytest.mark.bgpd, pytest.mark.bfdd]
+
+ROUTES = 20000
+
+# r1 neighbor -> r2 neighbor, multihop
+SESSIONS = {
+ "10.254.0.2": ("10.254.0.1", True),
+ "10.0.12.2": ("10.0.12.1", False),
+}
+
+
+def build_topo(tgen):
+ tgen.add_router("r1")
+ tgen.add_router("r2")
+
+ switch = tgen.add_switch("s1")
+ switch.add_link(tgen.gears["r1"])
+ switch.add_link(tgen.gears["r2"])
+
+
+def setup_module(mod):
+ tgen = Topogen(build_topo, mod.__name__)
+ tgen.start_topology()
+
+ for rname, router in tgen.routers().items():
+ router.load_frr_config(
+ os.path.join(CWD, "{}/frr.conf".format(rname)),
+ [
+ (TopoRouter.RD_ZEBRA, None),
+ (TopoRouter.RD_MGMTD, None),
+ (TopoRouter.RD_STATIC, None),
+ (TopoRouter.RD_BFD, None),
+ (TopoRouter.RD_BGP, None),
+ (TopoRouter.RD_SHARP, None),
+ ],
+ )
+
+ tgen.start_router()
+
+
+def teardown_module(mod):
+ tgen = get_topogen()
+ tgen.stop_topology()
+
+
+def _bfd_peer(router, peer):
+ for entry in json.loads(router.vtysh_cmd("show bfd peers json")):
+ if entry.get("peer") == peer:
+ return entry
+ return None
+
+
+def _check_converged(groups):
+ tgen = get_topogen()
+ r1 = tgen.gears["r1"]
+ r2 = tgen.gears["r2"]
+
+ summary = json.loads(r1.vtysh_cmd("show bgp ipv4 unicast summary json"))
+ for peer in SESSIONS:
+ info = summary.get("peers", {}).get(peer, {})
+ if info.get("state") != "Established":
+ return "r1: {} is {}".format(peer, info.get("state"))
+ if info.get("pfxRcd") != ROUTES:
+ return "r1: {} pfxRcd {}".format(peer, info.get("pfxRcd"))
+
+ for peer, (r2_peer, multihop) in SESSIONS.items():
+ entry = _bfd_peer(r1, peer)
+ if entry is None:
+ return "r1: bfdd has no session for {}".format(peer)
+ if entry.get("status") != "up" or entry.get("multihop") != multihop:
+ return "r1: bfdd session {}: {}".format(peer, entry)
+
+ entry = _bfd_peer(r2, r2_peer)
+ if entry is None or entry.get("status") != "up":
+ return "r2: bfdd session {}: {}".format(r2_peer, entry)
+
+ neigh = json.loads(r1.vtysh_cmd("show bgp neighbors {} json".format(peer)))
+ neigh = neigh.get(peer, {})
+ if neigh.get("peerGroup") != groups[peer]:
+ return "r1: {} peer-group {}, expected {}".format(
+ peer, neigh.get("peerGroup"), groups[peer]
+ )
+ bfd = neigh.get("peerBfdInfo", {})
+ if bfd.get("status") != "Up":
+ return "r1: bgpd BFD status for {}: {}".format(peer, bfd)
+
+ neigh = json.loads(r2.vtysh_cmd("show bgp neighbors {} json".format(r2_peer)))
+ bfd = neigh.get(r2_peer, {}).get("peerBfdInfo", {})
+ if bfd.get("status") != "Up":
+ return "r2: bgpd BFD status for {}: {}".format(r2_peer, bfd)
+
+ return None
+
+
+def _expect_converged(msg, groups):
+ test_func = functools.partial(_check_converged, groups)
+ _, result = topotest.run_and_expect(test_func, None, count=90, wait=1)
+ assert result is None, "{}: {}".format(msg, result)
+
+
+def _bgp_peer_count():
+ output = get_topogen().gears["r1"].vtysh_cmd("show memory", daemon="bgpd")
+ match = re.search(r"^BGP peer\s*:\s*(\d+)", output, re.M)
+ return int(match.group(1)) if match else None
+
+
+def _expect_stable(msg, groups, peers):
+ """
+ The old peer object is freed only after its routes are cleared: wait
+ until r1 bgpd is back to the number of peer objects it had before the
+ change, then check the session survived it.
+ """
+ _expect_converged(msg, groups)
+ _, result = topotest.run_and_expect(_bgp_peer_count, peers, count=60, wait=1)
+ assert result == peers, "{}: r1 bgpd has {} peer objects, expected {}".format(
+ msg, result, peers
+ )
+ time.sleep(2)
+ result = _check_converged(groups)
+ assert result is None, "{} (after old peer freed): {}".format(msg, result)
+
+
+def _r1_config(lines):
+ """Apply lines under router bgp on r1 and return the prior peer count."""
+ peers = _bgp_peer_count()
+ assert peers, "r1: cannot read the BGP peer count from show memory"
+ r1 = get_topogen().gears["r1"]
+ r1.vtysh_cmd("\n".join(["configure terminal", "router bgp 65001"] + lines))
+ return peers
+
+
+def test_bgp_bfd_converge():
+ tgen = get_topogen()
+ if tgen.routers_have_failure():
+ pytest.skip(tgen.errors)
+
+ step("Install {} routes on r2 and wait for BGP and BFD".format(ROUTES))
+ tgen.gears["r2"].vtysh_cmd(
+ "sharp install routes 10.100.0.0 nexthop 10.0.12.1 {}".format(ROUTES)
+ )
+ _expect_converged(
+ "initial convergence", {"10.254.0.2": "EDGE1", "10.0.12.2": "SH1"}
+ )
+
+
+def test_bgp_bfd_peer_group_move():
+ tgen = get_topogen()
+ if tgen.routers_have_failure():
+ pytest.skip(tgen.errors)
+
+ step("Move neighbors EDGE1 -> EDGE2 and SH1 -> SH2")
+ peers = _r1_config(
+ [
+ "no neighbor 10.254.0.2 peer-group EDGE1",
+ "neighbor 10.254.0.2 peer-group EDGE2",
+ "no neighbor 10.0.12.2 peer-group SH1",
+ "neighbor 10.0.12.2 peer-group SH2",
+ ]
+ )
+ _expect_stable(
+ "after moving to EDGE2/SH2", {"10.254.0.2": "EDGE2", "10.0.12.2": "SH2"}, peers
+ )
+
+ step("Move neighbors back EDGE2 -> EDGE1 and SH2 -> SH1")
+ peers = _r1_config(
+ [
+ "no neighbor 10.254.0.2 peer-group EDGE2",
+ "neighbor 10.254.0.2 peer-group EDGE1",
+ "no neighbor 10.0.12.2 peer-group SH2",
+ "neighbor 10.0.12.2 peer-group SH1",
+ ]
+ )
+ _expect_stable(
+ "after moving back to EDGE1/SH1",
+ {"10.254.0.2": "EDGE1", "10.0.12.2": "SH1"},
+ peers,
+ )
+
+
+def test_bgp_bfd_neighbor_delete_readd():
+ tgen = get_topogen()
+ if tgen.routers_have_failure():
+ pytest.skip(tgen.errors)
+
+ step("Delete neighbors and re-add them into EDGE2/SH2")
+ peers = _r1_config(
+ [
+ "no neighbor 10.254.0.2",
+ "neighbor 10.254.0.2 peer-group EDGE2",
+ "no neighbor 10.0.12.2",
+ "neighbor 10.0.12.2 peer-group SH2",
+ ]
+ )
+ _expect_stable(
+ "after delete and re-add", {"10.254.0.2": "EDGE2", "10.0.12.2": "SH2"}, peers
+ )
+
+
+def test_memory_leak():
+ "Run the memory leak test and report results."
+ tgen = get_topogen()
+ if not tgen.is_memleak_enabled():
+ pytest.skip("Memory leak test/report is disabled")
+
+ tgen.report_memory_leaks()
+
+
+if __name__ == "__main__":
+ args = ["-s"] + sys.argv[1:]
+ sys.exit(pytest.main(args))