1
0
mirror of https://github.com/home-assistant/operating-system.git synced 2025-12-19 18:08:29 +00:00

Fix broken outgoing connections caused by upstream kernel regression (#4267)

Upstream commit [1] caused regression in IPv4 routing which can cause some
routes becoming broadcast even though they should be routed as unicast, e.g.:

 # ip route get 1.1.1.1
 broadcast 1.1.1.1 via 192.168.122.1 dev enp0s3 src 192.168.122.204 uid 0
     cache <local,brd>

It's not entirely clear yet why it happens but this behavior seems to be
triggered for instance when the SSDP integration sends the broadcast packet on
HA startup. While this behavior is not described in the regression report [1],
the commit cherry-picked from Linux master fixes the problems for us as well.

Patches moved to version-specific folder, as this one shouldn't be applied on
Raspberry Pi targets.

[1] https://lore.kernel.org/all/20250710142714.12986-1-oscmaes92@gmail.com/
[2] https://lore.kernel.org/stable/20250822165231.4353-4-bacs@librecast.net/

Fixes #4265
This commit is contained in:
Jan Čermák
2025-09-01 18:16:11 +02:00
committed by GitHub
parent 44108a5493
commit 78bda4bd10
2 changed files with 103 additions and 0 deletions

View File

@@ -0,0 +1,52 @@
From 3a369be3f99ff577e09f8fc02703c7c9e4e74f3d Mon Sep 17 00:00:00 2001
From: Stefan Agner <stefan@agner.ch>
Date: Tue, 28 Mar 2023 12:02:10 +0200
Subject: [PATCH] ipv6: add option to explicitly enable reachability test
Systems which act as host as well as router might prefer the host
behavior. Currently the kernel does not allow to use IPv6 forwarding
globally and at the same time use route reachability probing.
Add a compile time flag to enable route reachability probe in any
case.
Signed-off-by: Stefan Agner <stefan@agner.ch>
---
net/ipv6/Kconfig | 9 +++++++++
net/ipv6/route.c | 3 ++-
2 files changed, 11 insertions(+), 1 deletion(-)
diff --git a/net/ipv6/Kconfig b/net/ipv6/Kconfig
index 1c9c686d9522f..ee4d9ca9d2e24 100644
--- a/net/ipv6/Kconfig
+++ b/net/ipv6/Kconfig
@@ -48,6 +48,15 @@ config IPV6_OPTIMISTIC_DAD
If unsure, say N.
+config IPV6_REACHABILITY_PROBE
+ bool "IPv6: Always use reachability probing (RFC 4191)"
+ help
+ By default reachability probing is disabled on router devices (when
+ IPv6 forwarding is enabled). This option explicitly enables
+ reachability probing always.
+
+ If unsure, say N.
+
config INET6_AH
tristate "IPv6: AH transformation"
select XFRM_AH
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 8ebfed5d63232..f1a61af0f5199 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -2223,7 +2223,8 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
strict |= flags & RT6_LOOKUP_F_IFACE;
strict |= flags & RT6_LOOKUP_F_IGNORE_LINKSTATE;
- if (READ_ONCE(net->ipv6.devconf_all->forwarding) == 0)
+ if (READ_ONCE(net->ipv6.devconf_all->forwarding) == 0 ||
+ IS_ENABLED(CONFIG_IPV6_REACHABILITY_PROBE))
strict |= RT6_LOOKUP_F_REACHABLE;
rcu_read_lock();

View File

@@ -0,0 +1,51 @@
From f356e2a2604bfc26be8bc51bf83f6a023f06316f Mon Sep 17 00:00:00 2001
From: Oscar Maes <oscmaes92@gmail.com>
Date: Wed, 27 Aug 2025 08:23:21 +0200
Subject: [PATCH] net: ipv4: fix regression in local-broadcast routes
[ Upstream commit 5189446ba995556eaa3755a6e875bc06675b88bd ]
Commit 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
introduced a regression where local-broadcast packets would have their
gateway set in __mkroute_output, which was caused by fi = NULL being
removed.
Fix this by resetting the fib_info for local-broadcast packets. This
preserves the intended changes for directed-broadcast packets.
Cc: stable@vger.kernel.org
Fixes: 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
Reported-by: Brett A C Sheffield <bacs@librecast.net>
Closes: https://lore.kernel.org/regressions/20250822165231.4353-4-bacs@librecast.net
Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://patch.msgid.link/20250827062322.4807-1-oscmaes92@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
net/ipv4/route.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 9a5c9497b3931..261ddb6542a40 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2532,12 +2532,16 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
!netif_is_l3_master(dev_out))
return ERR_PTR(-EINVAL);
- if (ipv4_is_lbcast(fl4->daddr))
+ if (ipv4_is_lbcast(fl4->daddr)) {
type = RTN_BROADCAST;
- else if (ipv4_is_multicast(fl4->daddr))
+
+ /* reset fi to prevent gateway resolution */
+ fi = NULL;
+ } else if (ipv4_is_multicast(fl4->daddr)) {
type = RTN_MULTICAST;
- else if (ipv4_is_zeronet(fl4->daddr))
+ } else if (ipv4_is_zeronet(fl4->daddr)) {
return ERR_PTR(-EINVAL);
+ }
if (dev_out->flags & IFF_LOOPBACK)
flags |= RTCF_LOCAL;