1. Patchew
  2. linux
  3. View series

[PATCH net] udp: clear skb->dev before running a sockmap verdict

Sechang Lim posted 1 patch 4 days, 18 hours ago
There is a newer version of this series
net/ipv4/udp.c | 5 +++++
1 file changed, 5 insertions(+)
[PATCH net] udp: clear skb->dev before running a sockmap verdict
Posted by Sechang Lim 4 days, 18 hours ago 
On the UDP receive path skb->dev is repurposed as dev_scratch (the
truesize/state cache set by udp_set_dev_scratch()), through the
union { struct net_device *dev; unsigned long dev_scratch; } in sk_buff.

When a UDP socket is in a sockmap, sk_data_ready is
sk_psock_verdict_data_ready(), which calls udp_read_skb() -> recv_actor()
(sk_psock_verdict_recv) to run the attached SK_SKB verdict program in softirq.
If that program calls a socket-lookup helper (bpf_sk_lookup_tcp/udp,
bpf_skc_lookup_tcp), bpf_skc_lookup() does:

	if (skb->dev)
		caller_net = dev_net(skb->dev);

skb->dev still holds the dev_scratch value (a non-NULL integer), so dev_net()
dereferences it as a struct net_device * and the kernel takes a general
protection fault on a non-canonical address in softirq:

  Oops: general protection fault, probably for non-canonical address 0x1010000800004a0
  CPU: 1 UID: 0 PID: 1406 Comm: syz.2.19 Not tainted 7.1.0-rc6 #1 PREEMPT(full)
  RIP: 0010:bpf_skc_lookup net/core/filter.c:7033 [inline]
  RIP: 0010:bpf_sk_lookup+0x45/0x160 net/core/filter.c:7047
  Call Trace:
   <IRQ>
   bpf_prog_4675cb904b7071f8+0x12e/0x14e
   bpf_prog_run_pin_on_cpu+0xc6/0x1f0
   sk_psock_verdict_recv+0x1ba/0x350
   udp_read_skb+0x31a/0x370
   sk_psock_verdict_data_ready+0x2e3/0x600
   __udp_enqueue_schedule_skb+0x4c8/0x650
   udpv6_queue_rcv_one_skb+0x3ec/0x740
   udp6_unicast_rcv_skb+0x11d/0x140
   ip6_protocol_deliver_rcu+0x61e/0x950
   ip6_input_finish+0xa9/0x150
   NF_HOOK+0x286/0x2f0
   ip6_input+0x117/0x220
   NF_HOOK+0x286/0x2f0
   __netif_receive_skb+0x85/0x200
   process_backlog+0x374/0x9a0
   __napi_poll+0x4f/0x1c0
   net_rx_action+0x3b0/0x770
   handle_softirqs+0x15a/0x460
   do_softirq+0x57/0x80
   </IRQ>

The rmem charge that dev_scratch accounted for is released by skb_recv_udp() on
dequeue, just above, so the scratch is dead by the time recv_actor() runs. Clear
skb->dev so bpf_skc_lookup() falls back to sock_net(skb->sk), which
skb_set_owner_sk_safe() set just above.

Fixes: 965b57b469a5 ("net: Introduce a new proto_ops ->read_skb()")
Signed-off-by: Sechang Lim <rhkrqnwk98@gmail.com>
---
 net/ipv4/udp.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 0ac2bf4f8759..f5ba36e721cb 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -2011,6 +2011,11 @@ int udp_read_skb(struct sock *sk, skb_read_actor_t recv_actor)
 	}
 
 	WARN_ON_ONCE(!skb_set_owner_sk_safe(skb, sk));
+	/* skb->dev still aliases the UDP rx dev_scratch (its charge was freed
+	 * on dequeue above); a sockmap verdict program may deref it via
+	 * bpf_sk_lookup_*(), so clear it -> bpf_skc_lookup() uses skb->sk
+	 */
+	skb->dev = NULL;
 	return recv_actor(sk, skb);
 }
 
-- 
2.43.0
Re: [PATCH net] udp: clear skb->dev before running a sockmap verdict
Posted by Alexander Lobakin 4 days, 16 hours ago 
From: Sechang Lim <rhkrqnwk98@gmail.com>
Date: Wed,  3 Jun 2026 12:26:37 +0000

> On the UDP receive path skb->dev is repurposed as dev_scratch (the
> truesize/state cache set by udp_set_dev_scratch()), through the
> union { struct net_device *dev; unsigned long dev_scratch; } in sk_buff.

[...]

> Fixes: 965b57b469a5 ("net: Introduce a new proto_ops ->read_skb()")

I think you need to add

Cc: stable@vger.kernel.org

here to mark it as a candidate for backporting.

> Signed-off-by: Sechang Lim <rhkrqnwk98@gmail.com>
> ---
>  net/ipv4/udp.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
> index 0ac2bf4f8759..f5ba36e721cb 100644
> --- a/net/ipv4/udp.c
> +++ b/net/ipv4/udp.c
> @@ -2011,6 +2011,11 @@ int udp_read_skb(struct sock *sk, skb_read_actor_t recv_actor)
>  	}
>  
>  	WARN_ON_ONCE(!skb_set_owner_sk_safe(skb, sk));
> +	/* skb->dev still aliases the UDP rx dev_scratch (its charge was freed

Pls favor generic block comment style

	/*
	 * skb->dev ...

> +	 * on dequeue above); a sockmap verdict program may deref it via
> +	 * bpf_sk_lookup_*(), so clear it -> bpf_skc_lookup() uses skb->sk
> +	 */
> +	skb->dev = NULL;
>  	return recv_actor(sk, skb);

I'd put newlines before and after the added piece for readability.

	WARN_ON_ONCE(...);

	/*
	 * skb->dev ...
	 * ...
	 */
	skb->dev = NULL;

	return recv_actor(sk, skb);

>  }
Thanks,
Olek
[PATCH net v2] udp: clear skb->dev before running a sockmap verdict
Posted by Sechang Lim 4 days, 14 hours ago 
On the UDP receive path skb->dev is repurposed as dev_scratch (the
truesize/state cache set by udp_set_dev_scratch()), through the
union { struct net_device *dev; unsigned long dev_scratch; } in sk_buff.

When a UDP socket is in a sockmap, sk_data_ready is
sk_psock_verdict_data_ready(), which calls udp_read_skb() -> recv_actor()
(sk_psock_verdict_recv) to run the attached SK_SKB verdict program in softirq.
If that program calls a socket-lookup helper (bpf_sk_lookup_tcp/udp,
bpf_skc_lookup_tcp), bpf_skc_lookup() does:

	if (skb->dev)
		caller_net = dev_net(skb->dev);

skb->dev still holds the dev_scratch value (a non-NULL integer), so dev_net()
dereferences it as a struct net_device * and the kernel takes a general
protection fault on a non-canonical address in softirq:

  Oops: general protection fault, probably for non-canonical address 0x1010000800004a0
  CPU: 1 UID: 0 PID: 1406 Comm: syz.2.19 Not tainted 7.1.0-rc6 #1 PREEMPT(full)
  RIP: 0010:bpf_skc_lookup net/core/filter.c:7033 [inline]
  RIP: 0010:bpf_sk_lookup+0x45/0x160 net/core/filter.c:7047
  Call Trace:
   <IRQ>
   bpf_prog_4675cb904b7071f8+0x12e/0x14e
   bpf_prog_run_pin_on_cpu+0xc6/0x1f0
   sk_psock_verdict_recv+0x1ba/0x350
   udp_read_skb+0x31a/0x370
   sk_psock_verdict_data_ready+0x2e3/0x600
   __udp_enqueue_schedule_skb+0x4c8/0x650
   udpv6_queue_rcv_one_skb+0x3ec/0x740
   udp6_unicast_rcv_skb+0x11d/0x140
   ip6_protocol_deliver_rcu+0x61e/0x950
   ip6_input_finish+0xa9/0x150
   NF_HOOK+0x286/0x2f0
   ip6_input+0x117/0x220
   NF_HOOK+0x286/0x2f0
   __netif_receive_skb+0x85/0x200
   process_backlog+0x374/0x9a0
   __napi_poll+0x4f/0x1c0
   net_rx_action+0x3b0/0x770
   handle_softirqs+0x15a/0x460
   do_softirq+0x57/0x80
   </IRQ>

The rmem charge that dev_scratch accounted for is released by skb_recv_udp() on
dequeue, just above, so the scratch is dead by the time recv_actor() runs. Clear
skb->dev so bpf_skc_lookup() falls back to sock_net(skb->sk), which
skb_set_owner_sk_safe() set just above.

Fixes: 965b57b469a5 ("net: Introduce a new proto_ops ->read_skb()")
Cc: stable@vger.kernel.org
Signed-off-by: Sechang Lim <rhkrqnwk98@gmail.com>
---
v2:
 - add blank lines around the added codes (Olek)
 - use generic block comment style (Olek)
 - Cc: stable

 net/ipv4/udp.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 0ac2bf4f8759..70f6cbd4ef73 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -2011,6 +2011,14 @@ int udp_read_skb(struct sock *sk, skb_read_actor_t recv_actor)
 	}
 
 	WARN_ON_ONCE(!skb_set_owner_sk_safe(skb, sk));
+
+	/*
+	 * skb->dev still aliases the UDP rx dev_scratch (its charge was freed
+	 * on dequeue above); a sockmap verdict program may deref it via
+	 * bpf_sk_lookup_*(), so clear it -> bpf_skc_lookup() uses skb->sk
+	 */
+	skb->dev = NULL;
+
 	return recv_actor(sk, skb);
 }
 
-- 
2.43.0

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.