[PATCH net v3] net: bridge: mcast: don't truncate the port group walk on teardown

[PATCH net v3] net: bridge: mcast: don't truncate the port group walk on teardown

[Jun Yang]

__br_multicast_disable_port_ctx() and br_multicast_del_port() walk
port->mglist with hlist_for_each_entry_safe(). However,
br_multicast_find_del_pg() can also delete other entries from the same
list through br_multicast_fwd_src_remove() or __fwd_del_star_excl().

If such an entry is the iterator's saved next node, hlist_del_init()
clears its ->next and terminates the walk early. The reproducer triggers
this in both teardown walks, leaving port groups in the bridge mdb with
a dangling ->key.port after del_nbp() frees the port:

  BUG: KASAN: slab-use-after-free in __mdb_fill_info+0x1191/0x1320
   __mdb_fill_info+0x1191/0x1320
   br_mdb_dump+0x594/0xe40
   rtnl_mdb_dump+0x1cf/0x5d0

Use hlist_del_init_rcu() to unlink the group while preserving ->next.
br_multicast_del_pg() and the teardown walks run under
br->multicast_lock. The GC worker must acquire the same lock before
detaching the group for destruction, so the node remains alive while
the walk uses the preserved pointer.

Preserving ->next means a walk can now reach a group that an earlier
iteration already deleted as a side effect. That group is off mp->ports,
so br_multicast_find_del_pg() would fall through its port scan and hit
the trailing WARN_ON(1). Skip such groups at the top of that helper: a
port group is put on port->mglist when it is created and only unlinked
when it is deleted, so hlist_unhashed() identifies exactly this case.

Fixes: b08123684bd5 ("net: bridge: mcast: install S,G entries automatically based on reports")
Cc: stable@vger.kernel.org
Suggested-by: Nikolay Aleksandrov <razor@blackwall.org>
Reported-by: TencentOS Corvus AI <corvus@tencent.com>
Assisted-by: tencentos-corvus-ai:kimi-k3
Signed-off-by: Jun Yang <junvyyang@tencent.com>
---
v3: skip already-unlinked groups at the top of br_multicast_find_del_pg()
    instead of in each teardown walk, per Nikolay's suggestion, so the
    preserved ->next no longer leads to the trailing WARN_ON(1).

v2: https://lore.kernel.org/netdev/20260826014200.362304-1-littleddfu@gmail.com/
    Use hlist_del_init_rcu() to preserve ->next, as suggested by Nikolay,
    instead of restarting the walks.

v1: https://lore.kernel.org/all/20260812113435.1854275-1-junvyyang@tencent.com/

 net/bridge/br_multicast.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 00aa9b2879d6..ea0a50eb8b77 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -808,7 +808,11 @@ void br_multicast_del_pg(struct net_bridge_mdb_entry *mp,
 	struct hlist_node *tmp;
 
 	rcu_assign_pointer(*pp, pg->next);
-	hlist_del_init(&pg->mglist);
+	/* Keep ->next (held under multicast_lock, freed later by the GC work):
+	 * a port->mglist teardown walk may have latched this node as its next,
+	 * and deleting other groups of the same port must not truncate it.
+	 */
+	hlist_del_init_rcu(&pg->mglist);
 	br_multicast_eht_clean_sets(pg);
 	hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
 		br_multicast_del_group_src(ent, false);
@@ -835,6 +839,13 @@ static void br_multicast_find_del_pg(struct net_bridge *br,
 	struct net_bridge_mdb_entry *mp;
 	struct net_bridge_port_group *p;
 
+	/* A teardown walk over port->mglist can reach a group that an earlier
+	 * iteration already deleted as a side effect. It is off mp->ports by
+	 * now, so skip it instead of falling through to the WARN_ON() below.
+	 */
+	if (hlist_unhashed(&pg->mglist))
+		return;
+
 	mp = br_mdb_ip_get(br, &pg->key.addr);
 	if (WARN_ON(!mp))
 		return;
-- 
2.43.7

Re: [PATCH net v3] net: bridge: mcast: don't truncate the port group walk on teardown

[Nikolay Aleksandrov]

On 31/08/2026 14:13, Jun Yang wrote:
> __br_multicast_disable_port_ctx() and br_multicast_del_port() walk
> port->mglist with hlist_for_each_entry_safe(). However,
> br_multicast_find_del_pg() can also delete other entries from the same
> list through br_multicast_fwd_src_remove() or __fwd_del_star_excl().
> 
> If such an entry is the iterator's saved next node, hlist_del_init()
> clears its ->next and terminates the walk early. The reproducer triggers
> this in both teardown walks, leaving port groups in the bridge mdb with
> a dangling ->key.port after del_nbp() frees the port:
> 
>    BUG: KASAN: slab-use-after-free in __mdb_fill_info+0x1191/0x1320
>     __mdb_fill_info+0x1191/0x1320
>     br_mdb_dump+0x594/0xe40
>     rtnl_mdb_dump+0x1cf/0x5d0
> 
> Use hlist_del_init_rcu() to unlink the group while preserving ->next.
> br_multicast_del_pg() and the teardown walks run under
> br->multicast_lock. The GC worker must acquire the same lock before
> detaching the group for destruction, so the node remains alive while
> the walk uses the preserved pointer.
> 
> Preserving ->next means a walk can now reach a group that an earlier
> iteration already deleted as a side effect. That group is off mp->ports,
> so br_multicast_find_del_pg() would fall through its port scan and hit
> the trailing WARN_ON(1). Skip such groups at the top of that helper: a
> port group is put on port->mglist when it is created and only unlinked
> when it is deleted, so hlist_unhashed() identifies exactly this case.
> 
> Fixes: b08123684bd5 ("net: bridge: mcast: install S,G entries automatically based on reports")
> Cc: stable@vger.kernel.org
> Suggested-by: Nikolay Aleksandrov <razor@blackwall.org>
> Reported-by: TencentOS Corvus AI <corvus@tencent.com>
> Assisted-by: tencentos-corvus-ai:kimi-k3
> Signed-off-by: Jun Yang <junvyyang@tencent.com>
> ---
> v3: skip already-unlinked groups at the top of br_multicast_find_del_pg()
>      instead of in each teardown walk, per Nikolay's suggestion, so the
>      preserved ->next no longer leads to the trailing WARN_ON(1).
> 
> v2: https://lore.kernel.org/netdev/20260826014200.362304-1-littleddfu@gmail.com/
>      Use hlist_del_init_rcu() to preserve ->next, as suggested by Nikolay,
>      instead of restarting the walks.
> 
> v1: https://lore.kernel.org/all/20260812113435.1854275-1-junvyyang@tencent.com/
> 
>   net/bridge/br_multicast.c | 13 ++++++++++++-
>   1 file changed, 12 insertions(+), 1 deletion(-)
> 
> diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
> index 00aa9b2879d6..ea0a50eb8b77 100644
> --- a/net/bridge/br_multicast.c
> +++ b/net/bridge/br_multicast.c
> @@ -808,7 +808,11 @@ void br_multicast_del_pg(struct net_bridge_mdb_entry *mp,
>   	struct hlist_node *tmp;
>   
>   	rcu_assign_pointer(*pp, pg->next);
> -	hlist_del_init(&pg->mglist);
> +	/* Keep ->next (held under multicast_lock, freed later by the GC work):
> +	 * a port->mglist teardown walk may have latched this node as its next,
> +	 * and deleting other groups of the same port must not truncate it.
> +	 */
> +	hlist_del_init_rcu(&pg->mglist);
>   	br_multicast_eht_clean_sets(pg);
>   	hlist_for_each_entry_safe(ent, tmp, &pg->src_list, node)
>   		br_multicast_del_group_src(ent, false);
> @@ -835,6 +839,13 @@ static void br_multicast_find_del_pg(struct net_bridge *br,
>   	struct net_bridge_mdb_entry *mp;
>   	struct net_bridge_port_group *p;
>   
> +	/* A teardown walk over port->mglist can reach a group that an earlier
> +	 * iteration already deleted as a side effect. It is off mp->ports by
> +	 * now, so skip it instead of falling through to the WARN_ON() below.
> +	 */
> +	if (hlist_unhashed(&pg->mglist))
> +		return;
> +
>   	mp = br_mdb_ip_get(br, &pg->key.addr);
>   	if (WARN_ON(!mp))
>   		return;


This should be fine, let's also see what sashiko thinks. :)

Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>

Re: [PATCH net v3] net: bridge: mcast: don't truncate the port group walk on teardown

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Mon, 31 Aug 2026 19:13:04 +0800 you wrote:
> __br_multicast_disable_port_ctx() and br_multicast_del_port() walk
> port->mglist with hlist_for_each_entry_safe(). However,
> br_multicast_find_del_pg() can also delete other entries from the same
> list through br_multicast_fwd_src_remove() or __fwd_del_star_excl().
> 
> If such an entry is the iterator's saved next node, hlist_del_init()
> clears its ->next and terminates the walk early. The reproducer triggers
> this in both teardown walks, leaving port groups in the bridge mdb with
> a dangling ->key.port after del_nbp() frees the port:
> 
> [...]

Here is the summary with links:
  - [net,v3] net: bridge: mcast: don't truncate the port group walk on teardown
    https://git.kernel.org/netdev/net/c/5a3f7a683aee

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html