netdev - Re: [PATCH v1 2/6] net: Add a generic udp_offload_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 23 Nov 2015 22:37:35 -0800
From:	Alexander Duyck <alexander.duyck@...il.com>
To:	Anjali Singhai Jain <anjali.singhai@...el.com>,
	netdev@...r.kernel.org
Cc:	jesse@...nel.org, Kiran Patil <kiran.patil@...el.com>
Subject: Re: [PATCH v1 2/6] net: Add a generic udp_offload_get_port function

On 11/23/2015 01:02 PM, Anjali Singhai Jain wrote:
> The new function udp_offload_get_port replaces vxlan_get_rx_port().
> This is a generic function that will help replay all udp tunnel ports
> irrespective of tunnel type.
> This way when new udp tunnels get added this function need not change.
>
> Note: Drivers besides i40e are compile tested with this change.
>
> Signed-off-by: Anjali Singhai Jain <anjali.singhai@...el.com>
> Signed-off-by: Kiran Patil <kiran.patil@...el.com>
> ---

[...]

> diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
> index f938616..8597020 100644
> --- a/net/ipv4/udp_offload.c
> +++ b/net/ipv4/udp_offload.c
> @@ -290,6 +290,33 @@ unlock:
>   }
>   EXPORT_SYMBOL(udp_del_offload);
>
> +void udp_offload_get_port(struct net_device *dev)
> +{
> +	struct udp_offload_priv __rcu **head;
> +	struct udp_offload_priv *uo_priv;
> +	struct udp_offload *uo;
> +
> +	if (udp_offload_base)
> +		head = &udp_offload_base;
> +	else
> +		return;
> +
> +	spin_lock(&udp_offload_lock);
> +	uo_priv = udp_deref_protected(*head);
> +	for (; uo_priv != NULL; uo_priv = udp_deref_protected(*head)) {
> +		/* call the right add port */
> +		uo = uo_priv->offload;
> +		if (uo && dev->netdev_ops->ndo_add_udp_tunnel_port)
> +			dev->netdev_ops->ndo_add_udp_tunnel_port(dev,
> +							uo->family,
> +							uo->port,
> +							uo->tunnel_type);
> +		head = &uo_priv->next;
> +	}
> +	spin_unlock(&udp_offload_lock);
> +}
> +EXPORT_SYMBOL(udp_offload_get_port);
> +
>   struct sk_buff **udp_gro_receive(struct sk_buff **head, struct sk_buff *skb,
>   				 struct udphdr *uh)
>   {
>

So when I got to patch 5 I realized this approach is horribly broken for 
IPv6 tunnels.  The udp_offload_base is only populated if the family is 
AF_INET.  What do you guys plan to do to get support for AF_INET6?

You probably ought to look at something like what ended up being done 
for the IOAT stuff.  What you end up needing is to support the drivers 
querying for what ports are active, and receiving notifications of 
tunnel updates, and the tunnel side that will register some 
functionality allowing the active ports for a given tunnel type to be 
queried.

- Alex
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html