vrf_init_module() ignores register_netdevice_notifier() errors and
continues module initialization, which can leave VRF loaded without its
netdev notifier registered.

Check the error and fail module initialization early.

Signed-off-by: Minhong He <heminhong@kylinos.cn>
---
 drivers/net/vrf.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
index 46209917ae4d..a0557a3a7026 100644
--- a/drivers/net/vrf.c
+++ b/drivers/net/vrf.c
@@ -1932,7 +1932,9 @@ static int __init vrf_init_module(void)
 {
 	int rc;
 
-	register_netdevice_notifier(&vrf_notifier_block);
+	rc = register_netdevice_notifier(&vrf_notifier_block);
+	if (rc < 0)
+		return rc;
 
 	rc = register_pernet_subsys(&vrf_net_ops);
 	if (rc < 0)
-- 
2.25.1

On 8/3/26 3:00 AM, Minhong He wrote:
> vrf_init_module() ignores register_netdevice_notifier() errors and
> continues module initialization, which can leave VRF loaded without its
> netdev notifier registered.
> 
> Check the error and fail module initialization early.
> 

Fixes: 193125dbd8eb ("net: Introduce VRF device driver")

> Signed-off-by: Minhong He <heminhong@kylinos.cn>
> ---
>  drivers/net/vrf.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
> index 46209917ae4d..a0557a3a7026 100644
> --- a/drivers/net/vrf.c
> +++ b/drivers/net/vrf.c
> @@ -1932,7 +1932,9 @@ static int __init vrf_init_module(void)
>  {
>  	int rc;
>  
> -	register_netdevice_notifier(&vrf_notifier_block);
> +	rc = register_netdevice_notifier(&vrf_notifier_block);
> +	if (rc < 0)
> +		return rc;
>  
>  	rc = register_pernet_subsys(&vrf_net_ops);
>  	if (rc < 0)

Reviewed-by: David Ahern <dsahern@kernel.org>

On Mon, 3 Aug 2026 10:59:47 -0600 David Ahern wrote:
> Fixes: 193125dbd8eb ("net: Introduce VRF device driver")

Thanks for the tag and the review!

In this case I'm going to skip and take this via net-next
'cause register_netdevice_notifier() is guaranteed not
to fail today.

On 8/4/26 7:55 PM, Jakub Kicinski wrote:
> In this case I'm going to skip and take this via net-next
> 'cause register_netdevice_notifier() is guaranteed not
> to fail today.

considered that response, so it works for me.