[PATCH 1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()

[PATCH 1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()

[Dongli Zhang]

The commit 4f61f133f354 ("net: tap: NULL pointer derefence in
dev_parse_header_protocol when skb->dev is null") fixed a crash in
tap_get_user() by assigning skb->dev before calling tun_vnet_hdr_to_skb().
This is required because virtio_net_hdr_to_skb() may invoke
dev_parse_header_protocol(), which dereferences skb->dev. Without the
assignment, a NULL pointer dereference can occur.

However, tap_get_user_xdp() still parses the virtio-net header before
assigning skb->dev. When the vhost TX path passes an XDP buffer containing
a GSO virtio-net header but the protocol is set to zero on purpose,
tun_vnet_hdr_to_skb() can reach dev_parse_header_protocol() while skb->dev
is still NULL, resulting in a crash.

Fix this by looking up the tap device and assigning skb->dev before calling
tun_vnet_hdr_to_skb(), matching the ordering already used in
tap_get_user(). Preserve the existing RCU read-side critical section across
dev_queue_xmit().

Fixes: 924a9bc362a5 ("net: check if protocol extracted by virtio_net_hdr_set_proto is correct")
Cc: stable@vger.kernel.org
Assisted-by: Codex:GPT-5.5
Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>
---
 drivers/net/tap.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/drivers/net/tap.c b/drivers/net/tap.c
index fae115915c8e..5d2d34d24ce8 100644
--- a/drivers/net/tap.c
+++ b/drivers/net/tap.c
@@ -1074,10 +1074,21 @@ static int tap_get_user_xdp(struct tap_queue *q, struct xdp_buff *xdp)
 	skb_reset_mac_header(skb);
 	skb->protocol = eth_hdr(skb)->h_proto;
 
+	rcu_read_lock();
+	tap = rcu_dereference(q->tap);
+	if (!tap) {
+		kfree_skb(skb);
+		rcu_read_unlock();
+		return 0;
+	}
+	skb->dev = tap->dev;
+
 	if (vnet_hdr_len) {
 		err = tun_vnet_hdr_to_skb(q->flags, skb, gso);
-		if (err)
+		if (err) {
+			rcu_read_unlock();
 			goto err_kfree;
+		}
 	}
 
 	/* Move network header to the right position for VLAN tagged packets */
@@ -1085,15 +1096,8 @@ static int tap_get_user_xdp(struct tap_queue *q, struct xdp_buff *xdp)
 	    vlan_get_protocol_and_depth(skb, skb->protocol, &depth) != 0)
 		skb_set_network_header(skb, depth);
 
-	rcu_read_lock();
-	tap = rcu_dereference(q->tap);
-	if (tap) {
-		skb->dev = tap->dev;
-		skb_probe_transport_header(skb);
-		dev_queue_xmit(skb);
-	} else {
-		kfree_skb(skb);
-	}
+	skb_probe_transport_header(skb);
+	dev_queue_xmit(skb);
 	rcu_read_unlock();
 
 	return 0;
-- 
2.43.5

Re: [PATCH 1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()

[Willem de Bruijn]

Dongli Zhang wrote:
> The commit 4f61f133f354 ("net: tap: NULL pointer derefence in
> dev_parse_header_protocol when skb->dev is null") fixed a crash in
> tap_get_user() by assigning skb->dev before calling tun_vnet_hdr_to_skb().
> This is required because virtio_net_hdr_to_skb() may invoke
> dev_parse_header_protocol(), which dereferences skb->dev. Without the
> assignment, a NULL pointer dereference can occur.
> 
> However, tap_get_user_xdp() still parses the virtio-net header before
> assigning skb->dev. When the vhost TX path passes an XDP buffer containing
> a GSO virtio-net header but the protocol is set to zero on purpose,
> tun_vnet_hdr_to_skb() can reach dev_parse_header_protocol() while skb->dev
> is still NULL, resulting in a crash.
> 
> Fix this by looking up the tap device and assigning skb->dev before calling
> tun_vnet_hdr_to_skb(), matching the ordering already used in
> tap_get_user(). Preserve the existing RCU read-side critical section across
> dev_queue_xmit().
> 
> Fixes: 924a9bc362a5 ("net: check if protocol extracted by virtio_net_hdr_set_proto is correct")
> Cc: stable@vger.kernel.org
> Assisted-by: Codex:GPT-5.5
> Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>

Reviewed-by: Willem de Bruijn <willemb@google.com>

Re: [PATCH 1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()

[Michael S. Tsirkin]

On Sun, Aug 02, 2026 at 03:46:12PM -0700, Dongli Zhang wrote:
> The commit 4f61f133f354 ("net: tap: NULL pointer derefence in
> dev_parse_header_protocol when skb->dev is null") fixed a crash in
> tap_get_user() by assigning skb->dev before calling tun_vnet_hdr_to_skb().
> This is required because virtio_net_hdr_to_skb() may invoke
> dev_parse_header_protocol(), which dereferences skb->dev. Without the
> assignment, a NULL pointer dereference can occur.
> 
> However, tap_get_user_xdp() still parses the virtio-net header before
> assigning skb->dev. When the vhost TX path passes an XDP buffer containing
> a GSO virtio-net header but the protocol is set to zero on purpose,
> tun_vnet_hdr_to_skb() can reach dev_parse_header_protocol() while skb->dev
> is still NULL, resulting in a crash.
> 
> Fix this by looking up the tap device and assigning skb->dev before calling
> tun_vnet_hdr_to_skb(), matching the ordering already used in
> tap_get_user(). Preserve the existing RCU read-side critical section across
> dev_queue_xmit().
> 
> Fixes: 924a9bc362a5 ("net: check if protocol extracted by virtio_net_hdr_set_proto is correct")
> Cc: stable@vger.kernel.org
> Assisted-by: Codex:GPT-5.5
> Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>

Acked-by: Michael S. Tsirkin <mst@redhat.com>

> ---
>  drivers/net/tap.c | 24 ++++++++++++++----------
>  1 file changed, 14 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/net/tap.c b/drivers/net/tap.c
> index fae115915c8e..5d2d34d24ce8 100644
> --- a/drivers/net/tap.c
> +++ b/drivers/net/tap.c
> @@ -1074,10 +1074,21 @@ static int tap_get_user_xdp(struct tap_queue *q, struct xdp_buff *xdp)
>  	skb_reset_mac_header(skb);
>  	skb->protocol = eth_hdr(skb)->h_proto;
>  
> +	rcu_read_lock();
> +	tap = rcu_dereference(q->tap);
> +	if (!tap) {
> +		kfree_skb(skb);
> +		rcu_read_unlock();
> +		return 0;
> +	}
> +	skb->dev = tap->dev;
> +
>  	if (vnet_hdr_len) {
>  		err = tun_vnet_hdr_to_skb(q->flags, skb, gso);
> -		if (err)
> +		if (err) {
> +			rcu_read_unlock();
>  			goto err_kfree;
> +		}
>  	}
>  
>  	/* Move network header to the right position for VLAN tagged packets */
> @@ -1085,15 +1096,8 @@ static int tap_get_user_xdp(struct tap_queue *q, struct xdp_buff *xdp)
>  	    vlan_get_protocol_and_depth(skb, skb->protocol, &depth) != 0)
>  		skb_set_network_header(skb, depth);
>  
> -	rcu_read_lock();
> -	tap = rcu_dereference(q->tap);
> -	if (tap) {
> -		skb->dev = tap->dev;
> -		skb_probe_transport_header(skb);
> -		dev_queue_xmit(skb);
> -	} else {
> -		kfree_skb(skb);
> -	}
> +	skb_probe_transport_header(skb);
> +	dev_queue_xmit(skb);
>  	rcu_read_unlock();
>  
>  	return 0;
> -- 
> 2.43.5

Re: [PATCH 1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()

[patchwork-bot+netdevbpf]

Hello:

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

On Sun,  2 Aug 2026 15:46:12 -0700 you wrote:
> The commit 4f61f133f354 ("net: tap: NULL pointer derefence in
> dev_parse_header_protocol when skb->dev is null") fixed a crash in
> tap_get_user() by assigning skb->dev before calling tun_vnet_hdr_to_skb().
> This is required because virtio_net_hdr_to_skb() may invoke
> dev_parse_header_protocol(), which dereferences skb->dev. Without the
> assignment, a NULL pointer dereference can occur.
> 
> [...]

Here is the summary with links:
  - [1/1] net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()
    https://git.kernel.org/netdev/net/c/3874892dd27d

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