Re: [PATCH net 2/2] ovpn: respect peer refcount in CMD_NEW_PEER error path

[Sabrina Dubroca]

2026-05-12, 05:19:13 +0100, David Carlier wrote:
> diff --git a/drivers/net/ovpn/netlink.c b/drivers/net/ovpn/netlink.c
> index 291e2e5bb450..4c66c1ec497e 100644
> --- a/drivers/net/ovpn/netlink.c
> +++ b/drivers/net/ovpn/netlink.c
> @@ -462,10 +462,12 @@ int ovpn_nl_peer_new_doit(struct sk_buff *skb, struct genl_info *info)
>  sock_release:
>  	ovpn_socket_release(peer);
>  peer_release:
> -	/* release right away because peer was not yet hashed, thus it is not
> -	 * used in any context
> +	/* For UDP, the peer is unreachable until added to the hashtables, so
> +	 * dropping the initial reference is enough. For TCP, the peer may be
> +	 * concurrently reachable via sk_user_data->peer until
> +	 * ovpn_socket_release() detaches; rely on the refcount.
>  	 */
> -	ovpn_peer_release(peer);
> +	ovpn_peer_put(peer);
>  
>  	return ret;
>  }

nit: after this change, ovpn_peer_release() is only used within
peer.c, and can become static.

I think it'd be nicer to do that now with a v2, but if Antonio wants
to pick this patch up directly, the fix LGTM:

Reviewed-by: Sabrina Dubroca <sd@xxxxxxxxxxxxxxx>

-- 
Sabrina