public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH net] mpls: add seqcount to protect the platform_label{,s} pair
@ 2026-03-23 23:25 Sabrina Dubroca
  2026-03-27  1:50 ` patchwork-bot+netdevbpf
  0 siblings, 1 reply; 2+ messages in thread
From: Sabrina Dubroca @ 2026-03-23 23:25 UTC (permalink / raw)
  To: netdev
  Cc: Sabrina Dubroca, Eric W. Biederman, Kuniyuki Iwashima,
	Guillaume Nault, Yuan Tan, Yifan Wu, Juefei Pu, Xin Liu

The RCU-protected codepaths (mpls_forward, mpls_dump_routes) can have
an inconsistent view of platform_labels vs platform_label in case of a
concurrent resize (resize_platform_label_table, under
platform_mutex). This can lead to OOB accesses.

This patch adds a seqcount, so that we get a consistent snapshot.

Note that mpls_label_ok is also susceptible to this, so the check
against RTA_DST in rtm_to_route_config, done outside platform_mutex,
is not sufficient. This value gets passed to mpls_label_ok once more
in both mpls_route_add and mpls_route_del, so there is no issue, but
that additional check must not be removed.

Reported-by: Yuan Tan <tanyuan98@outlook.com>
Reported-by: Yifan Wu <yifanwucs@gmail.com>
Reported-by: Juefei Pu <tomapufckgml@gmail.com>
Reported-by: Xin Liu <bird@lzu.edu.cn>
Fixes: 7720c01f3f590 ("mpls: Add a sysctl to control the size of the mpls label table")
Fixes: dde1b38e873c ("mpls: Convert mpls_dump_routes() to RCU.")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
---
 include/net/netns/mpls.h |  1 +
 net/mpls/af_mpls.c       | 29 +++++++++++++++++++++++++----
 2 files changed, 26 insertions(+), 4 deletions(-)

diff --git a/include/net/netns/mpls.h b/include/net/netns/mpls.h
index 6682e51513ef..2073cbac2afb 100644
--- a/include/net/netns/mpls.h
+++ b/include/net/netns/mpls.h
@@ -17,6 +17,7 @@ struct netns_mpls {
 	size_t platform_labels;
 	struct mpls_route __rcu * __rcu *platform_label;
 	struct mutex platform_mutex;
+	seqcount_mutex_t platform_label_seq;
 
 	struct ctl_table_header *ctl;
 };
diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c
index d5417688f69e..18d3da8ab384 100644
--- a/net/mpls/af_mpls.c
+++ b/net/mpls/af_mpls.c
@@ -83,14 +83,30 @@ static struct mpls_route *mpls_route_input(struct net *net, unsigned int index)
 	return mpls_dereference(net, platform_label[index]);
 }
 
+static struct mpls_route __rcu **mpls_platform_label_rcu(struct net *net, size_t *platform_labels)
+{
+	struct mpls_route __rcu **platform_label;
+	unsigned int sequence;
+
+	do {
+		sequence = read_seqcount_begin(&net->mpls.platform_label_seq);
+		platform_label = rcu_dereference(net->mpls.platform_label);
+		*platform_labels = net->mpls.platform_labels;
+	} while (read_seqcount_retry(&net->mpls.platform_label_seq, sequence));
+
+	return platform_label;
+}
+
 static struct mpls_route *mpls_route_input_rcu(struct net *net, unsigned int index)
 {
 	struct mpls_route __rcu **platform_label;
+	size_t platform_labels;
+
+	platform_label = mpls_platform_label_rcu(net, &platform_labels);
 
-	if (index >= net->mpls.platform_labels)
+	if (index >= platform_labels)
 		return NULL;
 
-	platform_label = rcu_dereference(net->mpls.platform_label);
 	return rcu_dereference(platform_label[index]);
 }
 
@@ -2240,8 +2256,7 @@ static int mpls_dump_routes(struct sk_buff *skb, struct netlink_callback *cb)
 	if (index < MPLS_LABEL_FIRST_UNRESERVED)
 		index = MPLS_LABEL_FIRST_UNRESERVED;
 
-	platform_label = rcu_dereference(net->mpls.platform_label);
-	platform_labels = net->mpls.platform_labels;
+	platform_label = mpls_platform_label_rcu(net, &platform_labels);
 
 	if (filter.filter_set)
 		flags |= NLM_F_DUMP_FILTERED;
@@ -2645,8 +2660,12 @@ static int resize_platform_label_table(struct net *net, size_t limit)
 	}
 
 	/* Update the global pointers */
+	local_bh_disable();
+	write_seqcount_begin(&net->mpls.platform_label_seq);
 	net->mpls.platform_labels = limit;
 	rcu_assign_pointer(net->mpls.platform_label, labels);
+	write_seqcount_end(&net->mpls.platform_label_seq);
+	local_bh_enable();
 
 	mutex_unlock(&net->mpls.platform_mutex);
 
@@ -2728,6 +2747,8 @@ static __net_init int mpls_net_init(struct net *net)
 	int i;
 
 	mutex_init(&net->mpls.platform_mutex);
+	seqcount_mutex_init(&net->mpls.platform_label_seq, &net->mpls.platform_mutex);
+
 	net->mpls.platform_labels = 0;
 	net->mpls.platform_label = NULL;
 	net->mpls.ip_ttl_propagate = 1;
-- 
2.51.2


^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-03-27  1:50 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-23 23:25 [PATCH net] mpls: add seqcount to protect the platform_label{,s} pair Sabrina Dubroca
2026-03-27  1:50 ` patchwork-bot+netdevbpf

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox