netdev - Re: [PATCH net-next V2 3/3] net: Add GRO support for vxlan traffic

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:	Tue, 7 Jan 2014 22:10:44 +0200
From:	Or Gerlitz <or.gerlitz@...il.com>
To:	Eric Dumazet <eric.dumazet@...il.com>
Cc:	Or Gerlitz <ogerlitz@...lanox.com>, Jerry Chu <hkchu@...gle.com>,
	Eric Dumazet <edumazet@...gle.com>,
	Herbert Xu <herbert@...dor.apana.org.au>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	David Miller <davem@...emloft.net>,
	Yan Burman <yanb@...lanox.com>,
	Shlomo Pongratz <shlomop@...lanox.com>
Subject: Re: [PATCH net-next V2 3/3] net: Add GRO support for vxlan traffic

On Tue, Jan 7, 2014 at 10:04 PM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> On Tue, 2014-01-07 at 21:43 +0200, Or Gerlitz wrote:
>> On Tue, Jan 7, 2014 at 6:34 PM, Eric Dumazet <eric.dumazet@...il.com> wrote:
>> > On Tue, 2014-01-07 at 17:29 +0200, Or Gerlitz wrote:
>> >> Add gro handlers for vxlan using the udp gro infrastructure
>> >>
>> >
>> >
>> >>  static void vxlan_notify_add_rx_port(struct sock *sk)
>> >>  {
>> >> @@ -568,6 +661,8 @@ static void vxlan_notify_add_rx_port(struct sock *sk)
>> >>                       dev->netdev_ops->ndo_add_vxlan_port(dev, sa_family,
>> >>                                                           port);
>> >>       }
>> >> +     if (sa_family == AF_INET)
>> >> +             udp_add_offload(&vxlan_offload, port);
>> >>       rcu_read_unlock();
>> >>  }
>> >>
>> >
>> > This means two vxlan tunnels can not share same port.
>> > Is that a valid assertion ?
>>
>> nope -- the vxlan driver opens a listener udp socket per per listening
>> port, but N > 1 vxlan tunnels can sit on that port, the driver does
>> further demuxing based on the vnid carried in the vxlan header, see
>> the call to vxlan_find_vni() from vxlan_rcv()
>> --
>
> So if we use same port, it seems to me udp_del_offload() of the first
> dismantled tunnel will remove the offload for the remaining tunnel.
>
> You forgot to implement a refcount somehow.

no, the vxlan driver does it for me on a higher level, you can see in
vxlan_sock_release() that vxlan_notify_del_rx_port() is called only if
the ref count which is associated with the listener socket drops to
zero.
--
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