[pve-devel] [PATCH frr 1/2] d/patches: add backport for bfd session down fix

* [pve-devel] [PATCH frr 1/2] d/patches: add backport for bfd session down fix
@ 2025-04-18 11:21 Stefan Hanreich
  2025-04-18 11:21 ` [pve-devel] [PATCH frr 2/2] bump version to 10.2.2-1+pve2 Stefan Hanreich
  2025-04-22  9:39 ` [pve-devel] [PATCH frr 1/2] d/patches: add backport for bfd session down fix Stefan Hanreich
  0 siblings, 2 replies; 3+ messages in thread
From: Stefan Hanreich @ 2025-04-18 11:21 UTC (permalink / raw)
  To: pve-devel

This is a backport of the FRR PR #18562 [1], that has been submitted
upstream as well [2].

With already established BGP/BFD sessions via underlay networks that
use a different routing protocol than BGP, rebooting a node could
cause the BFD session to enter the ADMIN_DOWN state without ever
recovering from it. This could be alleviated by manually resetting the
BFD session. The commit was identified by bisecting and manually
backported, since the upstream commit was on top of a refactored
bgpd.c file, so it didn't apply cleanly anymore.

The jump to version 10.2.2-1+pve1 fixed this issue for BGP underlay
networks, but the issue could still crop up with OpenFabric / OSPF /
IS-IS underlay networks.

[1] https://github.com/FRRouting/frr/pull/18562
[2] https://github.com/FRRouting/frr/pull/18682

Signed-off-by: Stefan Hanreich <s.hanreich@proxmox.com>
---
 ...eer-as-not-active-due-to-BFD-down-on.patch | 33 +++++++++++++++++++
 debian/patches/series                         |  1 +
 2 files changed, 34 insertions(+)
 create mode 100644 debian/patches/pve/0007-bgpd-Treat-the-peer-as-not-active-due-to-BFD-down-on.patch

diff --git a/debian/patches/pve/0007-bgpd-Treat-the-peer-as-not-active-due-to-BFD-down-on.patch b/debian/patches/pve/0007-bgpd-Treat-the-peer-as-not-active-due-to-BFD-down-on.patch
new file mode 100644
index 0000000..6d2c379
--- /dev/null
+++ b/debian/patches/pve/0007-bgpd-Treat-the-peer-as-not-active-due-to-BFD-down-on.patch
@@ -0,0 +1,33 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Stefan Hanreich <s.hanreich@proxmox.com>
+Date: Fri, 18 Apr 2025 10:36:19 +0200
+Subject: [PATCH] bgpd: Treat the peer as not active due to BFD down only if
+ established (backport #18562)
+
+If we have `neighbor X bfd` and BFD status is DOWN and/or ADMIN_DOWN, and BGP
+session is not yet established, we never allow the session to establish.
+
+Let's fix this regression that was in 10.2.
+
+Fixes: 1fb48f5 ("bgpd: Do not start BGP session if BFD profile is in shutdown state")
+Signed-off-by: Stefan Hanreich <s.hanreich@proxmox.com>
+---
+ bgpd/bgpd.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/bgpd/bgpd.c b/bgpd/bgpd.c
+index 0601400afd..f6b8abd9cc 100644
+--- a/bgpd/bgpd.c
++++ b/bgpd/bgpd.c
+@@ -4648,7 +4648,7 @@ bool peer_active(struct peer *peer)
+ 		return false;
+ 
+ 	if (peer->bfd_config) {
+-		if (bfd_session_is_down(peer->bfd_config->session))
++		if (peer_established(peer->connection) && bfd_session_is_down(peer->bfd_config->session))
+ 			return false;
+ 	}
+ 
+-- 
+2.39.5
+
diff --git a/debian/patches/series b/debian/patches/series
index c59ef31..e795d6c 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -4,3 +4,4 @@ pve/0003-tests-add-bgp-evpn-autort-test.patch
 pve/0004-zebra-add-ZEBRA_IF_DUMMY-flag-for-dummy-interfaces.patch
 pve/0005-fabricd-add-option-to-treat-dummy-interfaces-as-loop.patch
 pve/0006-fabricd-enable-dummy_as_loopback-option-per-default.patch
+pve/0007-bgpd-Treat-the-peer-as-not-active-due-to-BFD-down-on.patch
-- 
2.39.5


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel


^ permalink raw reply	[flat|nested] 3+ messages in thread