[PATCH net v2] ipv4: fix use-after-free in fib_nhc_update_mtu()

From: Chengfeng Ye

Date: Thu Aug 06 2026 - 09:34:27 EST


fib_nhc_update_mtu() walks the nexthop exception table under RTNL, but
RTNL does not serialize this walk with PMTU exception updates. The walk
uses rcu_dereference_protected() with a constant true condition without
holding fnhe_lock.

The following interleaving can therefore occur:

CPU 0 CPU 1
fib_nhc_update_mtu() update_or_create_fnhe()
load fnhe spin_lock_bh(&fnhe_lock)
fnhe_remove_oldest()
unlink fnhe
kfree_rcu(fnhe, rcu)
<quiescent state>
access fnhe after grace period

KASAN reported:

BUG: KASAN: slab-use-after-free in fib_nhc_update_mtu+0x3df/0x410
Read of size 8 at addr ffff888107d49000 by task poc/90
Call Trace:
fib_nhc_update_mtu+0x3df/0x410
fib_sync_mtu+0x7a/0xd0
fib_netdev_event+0x229/0x3f0
netif_set_mtu_ext+0x33a/0x570
dev_set_mtu+0x88/0x120
Allocated by task 89:
update_or_create_fnhe+0xa80/0x1110
__ip_rt_update_pmtu+0x8f2/0xcd0
ipv4_sk_update_pmtu+0x49e/0x690
udp_err+0xd92/0x1080
Freed by task 0:
__kasan_slab_free+0x43/0x70
kvfree_rcu_cb+0x12f/0x420
rcu_core+0x509/0x18e0

The same walk updates fnhe_pmtu and fnhe_mtu_locked. These fields form a
pair and other writers serialize them with fnhe_lock. RCU alone would
prevent reclamation, but would still allow concurrent writers to leave a
mixed pair.

Expose fnhe_lock to fib_semantics.c and hold it across the exception-table
walk. This prevents entries from being unlinked while they are visited and
serializes the paired PMTU state updates with all other writers.

Fixes: af7d6cce5369 ("net: ipv4: update fnhe_pmtu when first hop's MTU changes")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Chengfeng Ye <nicoyip.dev@xxxxxxxxx>
---
Changes in v2:
- Protect the walk with fnhe_lock rather than RCU alone, covering both
exception lifetime and the paired PMTU state updates.
- Keep fib_nhc_update_mtu() in fib_semantics.c and expose fnhe_lock for a
smaller diff.

Link: https://lore.kernel.org/netdev/20260731162938.3388534-1-nicoyip.dev@xxxxxxxxx/ [v1]

include/net/ip_fib.h | 4 ++++
net/ipv4/fib_semantics.c | 5 ++++-
net/ipv4/route.c | 2 +-
3 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index c63a3c4967ae..634a669ba46b 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -12,6 +12,7 @@
#ifndef _NET_IP_FIB_H
#define _NET_IP_FIB_H

+#include <linux/spinlock.h>
#include <net/flow.h>
#include <linux/seq_file.h>
#include <linux/rcupdate.h>
@@ -495,6 +496,9 @@ int fib_sync_up(struct net_device *dev, unsigned char nh_flags);
void fib_sync_mtu(struct net_device *dev, u32 orig_mtu);
void fib_nhc_update_mtu(struct fib_nh_common *nhc, u32 new, u32 orig);

+/* Protects nexthop exception table updates. */
+extern spinlock_t fnhe_lock;
+
/* Fields used for sysctl_fib_multipath_hash_fields.
* Common to IPv4 and IPv6.
*
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index 4f3c0740dde9..f091a8ac1974 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -1879,9 +1879,10 @@ void fib_nhc_update_mtu(struct fib_nh_common *nhc, u32 new, u32 orig)
struct fnhe_hash_bucket *bucket;
int i;

+ spin_lock_bh(&fnhe_lock);
bucket = rcu_dereference_protected(nhc->nhc_exceptions, 1);
if (!bucket)
- return;
+ goto out;

for (i = 0; i < FNHE_HASH_SIZE; i++) {
struct fib_nh_exception *fnhe;
@@ -1900,6 +1901,8 @@ void fib_nhc_update_mtu(struct fib_nh_common *nhc, u32 new, u32 orig)
}
}
}
+out:
+ spin_unlock_bh(&fnhe_lock);
}

void fib_sync_mtu(struct net_device *dev, u32 orig_mtu)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 152d8cb28f65..fc9a51ebd922 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -571,7 +571,7 @@ static void ip_rt_build_flow_key(struct flowi4 *fl4, const struct sock *sk,
build_sk_flow_key(fl4, sk);
}

-static DEFINE_SPINLOCK(fnhe_lock);
+DEFINE_SPINLOCK(fnhe_lock);

static void fnhe_flush_routes(struct fib_nh_exception *fnhe)
{
--
2.43.0